Tennessee
Valley
Authority
United Stales
Environmental Protection
Agency
Research and Development
Division of Environmental
Planning
Chattanooga TN 37401
TVA/EP-78/10
Office of Energy, Minerals, and  EPA-600/7-7B-154
Industry        August 1978
Washington DC 20460
Applications
of Computer Graphics
to Integrated
Environmental
Assessments
of Energy Systems

Interagency
Energy/Environment
R&D Program
Report
                ~z

-------
                 RESEARCH REPORTING SERIES

 Research reports of the Office of Research and Development, U.S. Environmental
 Protection Agency, have been grouped into nine series. These nine broad cate-
 gories were established to facilitate further development and application of en-
 vironmental technology. Elimination of traditional grouping was  consciously
 planned to foster technology transfer and a maximum interface in related fields.
 The nine series are:

       1.  Environmental Health Effects Research
      2.  Environmental Protection Technology
      3.  Ecological Research
      4.  Environmental Monitoring
      5.  Socioeconomic Environmental Studies
      6.  Scientific and Technical Assessment Reports (STAR)
      7.  Interagency Energy-Environment Research and Development
      8.  "Special" Reports
      9.  Miscellaneous Reports

 This report has been assigned to the  INTERAGENCY ENERGY-ENVIRONMENT
 RESEARCH AND DEVELOPMENT series. Reports in this series result from the
 effort funded under the 17-agency Federal Energy/Environment Research and
 Development Program. These studies relate to EPA's mission to protect the public
 health and welfare from adverse effects of pollutants associated with energy sys-
 tems. The goal of the Program is to assure the rapid development of domestic
 energy supplies in an environmentally-compatible manner by providing the nec-
 essary environmental data and control technology. Investigations include analy-
 ses of the transport  of energy-related pollutants and their health and ecological
 effects; assessments of, and  development of, control technologies for energy
 systems; and integrated assessments of a wide range of energy-related environ-
 mental issues.
This document is available to the public through the National Technical Informa-
tion Service. Springfield, Virginia 22161.

-------
                                             EPA-600/7-78-154
                                             TVA/EP-78/10
                                             August 1978
APPLICATIONS OF COMPUTER GRAPHICS TO INTEGRATED
  ENVIRONMENTAL ASSESSMENTS OF ENERGY SYSTEMS
                      by
  Malcolm C. Babb and Harrison R. Hickey, Jr.
      Division of Environmental Planning
          Tennessee Valley Authority
         Chattanooga, Tennessee  37401
       Interagency Agreement No. D5-E721
             Project No. E-AP 79BW
         Program Element No. INE 624C
                Project Officer
              Marvin Wayne Bloch
   Office of Energy, Minerals, and Industry
     U.S. Environmental Protection Agency
            Washington, D.C.  20460
                 Prepared for

   OFFICE OF ENERGY, MINERALS, AND INDUSTRY
      OFFICE OF RESEARCH AND DEVELOPMENT
     U.S. ENVIRONMENTAL PROTECTION AGENCY
            WASHINGTON, D.C.  20460

-------
                            DISCLAIMER
     This report was prepared by the Tennessee Valley Authority and
has been reviewed by the Office of Energy,  Minerals,  and Industry,  U.S.
Environmental Protection Agency, and approved for publication.   Approval
does not signify that the contents necessarily reflect the views and
policies of the Tennessee Valley Authority or the U.S. Environmental
Protection Agency, nor does mention of trade names or commercial products
constitute endorsement or recommendation for use.
                                11

-------
                             ABSTRACT
     This report summarizes the first two years of research designed to
demonstrate applications of computer graphics to environmental analyses
associated with the evaluation of impacts from development of conventional
energy systems.  The work emphasizes the use of storage-tube computer
graphics technology as a means for improving the interaction between the
engineer-scientist and the power of the computer.  Computer graphics is
also shown to be an effective medium for summarizing and communicating
information about the environment and pollution control alternatives to
technical specialists, managers, and the public.  The research has
resulted in a saving of time and cost for many analysis.  Also, many
techniques of analysis previously considered impractical can now be
conducted on a routine basis.  Applications to several fields of
analysis are described in detail, including air quality, water quality,
radiological hygiene, industrial hygiene, socioeconomics, and data
facilities siting with the use of geographically referenced data.
                                 111

-------
                             CONTENTS
Abstract	iii
List of Figures	vii
List of Tables	xi
Acknowledgments  	  xii

1.   Introduction	    1
          Integrated Assessment Research 	    1
          Evolution of Environmental Analysis Through the
            Use of Computer Techniques 	    1
          Interactive Computer Graphics	    3
2.   Conclusions 	    5
3.   Scope of Research	    6
          Research Objectives and Goals	    6
          Selection and Distribution of Computer Hardware
            and Software	    7
4.   Development of General-Purpose Graphics 	   12
          Interactive Graphics Display Package 	   12
          Kiviat Diagram Plotting Routine  	   18
          Interactive AG-II Plotting Routine 	   26
          Difficulties Involved in Implementing Graphics
            Hardware and Software	28
5.   Applications and Demonstrations of Computer Graphics	29
          Air Quality Analysis 	   29
            Display of Data	29
            Interactive Analysis of the Effects of Mobile
              Point Sources	32
            Analysis of Output from a Drift, Vapor, and Dry
              Plume Model	36
            Air Flow Behavior Patterns	42
            Future Demonstrations   	   42
          Water Quality Analysis 	   46
            Data Display	46
            Interactive Analysis of Water Quality Data Base. ...   49
            Water Source Mapping	49
            Interactive Analysis of Water Quality Models 	   57
            Future Applications	66
          Radiological Hygiene Data Display	67
          Industrial Hygiene—Display of Noise Data	67
          Socioeconomic Impact Analysis	73
          Geographic Information Systems as Aids to Siting
            Facilities	73
6.   Future Directions 	   98
7.   Summary	100
References	102

-------
                             CONTENTS
                            (continued)
Appendices
     A.   Summary of potential environmental analyses
            to which computer graphics could be
            applied	A-l
     B.   Difficulties with computer hardware and
            software encountered when implementing
            demonstrations of computer graphics	B-l
     C.   TVA Socioeconomic Impact Assessment Methods
            Project—development of site screening
            methods	C-l
                               VI

-------
                          LIST OF FIGURES
Number                                                            page

  1   Spectrum of currently available devices for computer
        graphics output 	    4

  2   Relationship between TVA's graphic software packages  ...   10

  3   Formatted tabulation of concentrations of dye in surface
        water, as generated by IGDP software	14

  4   Partitioned contour plot (ten intervals) of concentra-
        tions of dye in surface water, as generated by IGDP
        software	15

  5   Line contour plot of concentrations of dye in surface water,
        as generated by IGDP software	16

  6   Three-dimensional display of concentrations of dye in
        surface water, as generated by IDGP software  	   18

  7   User-selected profiles of dye concentrations perpendicular
        to the riverbank, as generated by IGDP software 	   19

  8   Butterfly configuration of a Kiviat diagram applied to the
        description of a socioeconomic system 	   21

  9   Star configuration of a Kiviat diagram depicting county-
        level educational indicators  	   24

 10   Nine examples of the use of the Kiviat diagram for
        engineering and management analyses  	   25

 11   Sample plots of data using the interactive AG-II plotting
        routine	27

 12   Profile of measured ground elevation and bedrock in the
        vicinity of soybean test plots	30

 13   Measured illumination levels at two elevations in a
       . vegetation growth chamber	31

 14   Display of air quality data in the vicinity of a steam
        plant	33

 15   Completed grid of mobile air pollutant source 	   34

 16   Three-dimensional display of downwind pollutant distri-
        bution pattern for a linear array of three gaseous
        emission sources	35
                                VI1

-------
Number                                                            Page

 17   Simulated behavior of a vapor plume from a natural-draft
        cooling tower, typical summer conditions	38

 18   Sample contours of yearly drift deposition in the vicinity
        of two natural-draft cooling towers 	   39

 19   Three-dimensional representation of yearly drift deposition
        in the vicinity of two natural-draft cooling towers  ...   40

 20   Plot of yearly drift deposition vs.  distance  from two
        natural-draft cooling towers	41

 21   Simulated pattern of air flow in the vicinity of a
        building	43

 22   Enlarged view of simulated  entrainment cavity on downwind
        side of building	44

 23   Various views of simulated  air flow  in the vicinity of  two
        mechanical-draft cooling  towers  	   45

 24   Typical display  of biological and physical water quality
        data  for TVA-EPA Thermal  Effects Studies	47

 25   Display of phytoplankton data collected as part of TVA-EPA
        Thermal Effects  Studies 	   48

 26   Three-dimensional  representation of  distribution of
        dissolved oxygen in the vicinity of a dam and a thermal
        discharge	49

 27   Three-dimensional  representation of  turbidity as a
        function of river mile and  depth	50

 28   Variation in  the National Science Foundation's water
        quality index  as a function of time and  Tennessee
        River reservoir	51

 29   Average weekly concentration  of dissolved  oxygen in
        discharge waters from Fort  Loudoun Reservoir over the
        past 16 years	53

 30   Yearly average concentration  of dissolved  oxygen in
        discharge waters from Fort  Loudoun Reservoir over
        the  past 16 years	54

 31   Yearly average amount of oxygen that would have to be
        added to discharge waters of Fort  Loudoun Reservoir
        for  the past 16  years to  bring the concentration
        to the specified levels	55
                               Vlll

-------
Number                                                            page

 32   Average temperatures of spring discharge water for
        reservoirs on the Tennessee River over the past
        16 years	56

 33   Initial display by cathode ray tube of water source
        symbols plotted on the basis of latitude and
        longitude	57

 34   Final display by cathode ray tube of water source symbols
        after visual adjustment 	   58

 35   Figure for final report prepared by interactive graphics
        showing water sources in the vicinity of a proposed
        nuclear electric generating facility	59

 36   Interactive scheme for the analysis of simulation
        results through the use of a mathematical model
        of temperature phenomena in a deep reservoir	63

 37   Profiles of predicted depth and temperature selected
        interactively by the user and displayed on a
        cathode ray tube	64

 38   Profile of predicted temperature and depth overlaid
        with actual temperature profile measured in the field  . .   65

 39   Plot of statistics (root mean square and average mean
        error) calculated over entire water column for a
        simulation run	66

 40   Dose conversion factor vs. uranium-238 particle size. ...   67

 41   Probability of a given radiation exposure based on
        field measurements	68

 42   Plot of time vs. distance for a given dose to thyroid
        gland for an accidental exposure situation	70

 43   Gamma-ray energy spectrum for cesium-144 as measured
        by a Ge(Li) detector	71

 44   Display of octave band noise measurements and predicted
        noise levels when personal protective equipment is
        used	72

 45   Display of county boundaries in the socioeconomic
        methodology test area	76

 46   Sample of socioeconomic  indicators displayed as Kiviat
        diagrams	77
                                IX

-------
Number                                                            Page

 47   Scatter plot of percent of population living in urban
        setting vs. average teacher's salary	78

 48   Shaded line-printer composite display of septic tank
        suitability for Knox County, Tennessee	83

 49   CRT display of cells classified as developed land in
        the demonstration area	86

 50   Overlay of major roads and 50-ft contours on CRT display
        of developed land	87

 51   CRT display of cells that possess various combinations  of
        (1) upland coniferous  forest,  (2)  a  land slope of 3 to
        6%, and (3) farms and  estates	88

 52   Calculated number of acres in the demonstration area that
        meet various  combinations  of the three characteristics
        selected for  analysis  	   89

 53   Sample of the computer graphics  used in  the Inter-
        divisional  Study Team  Report on the  use of geographic
        information by  TVA	95

-------
                          LIST OF TABLES

Number                                                            Page
 1   Distribution of Computer Graphics Hardware for Project
       Research	    9

 2   Numeric Rankings Among 22 Counties for Six Socioeconomic
       Indicators	22

 3   User-Defined Options for General-Purpose Plotting
       Package	26

 4   County-Level Site Screening Indicators Tested for Impact
       Analysis	73

 5   Catalog of Data Base Management, Analysis, and Graphics
       Routines for a County-Wide Soioeconomic Information
       System	74

 6   Spatial Data Classifications Used to Characterize Cells
       in the Demonstration Area	85

 7   Inventory of Geographic Information System Used by TVA's
       Division of Environmental Planning  	   90
                                 XI

-------
                         ACKNOWLEDGMENTS
     This work was conducted as part of the Federal Interagency Energy/
Environment Research and Development Program with funds administered
through the Environmental Protection Agency (EPA Contract No. EPA-IAG-
D5-E721, TVA Contract No. TV-41967A).

     The EPA Project Officer for this research project is Marvin Wayne
Bloch, Office of Energy, Minerals and Industry, U.S. Environmental Pro-
tection Agency, Washington, D.C.  The authors wish to acknowledge the
specific graphics contributions made to this work by Dr.  P.  R. Slawson
and W. J. McCormick, Envirodyne Limited, Waterloo, Ontario,  Canada;
Timothy L. Crawford and Dr. Norman Lacasse, Air Quality Branch, Muscle
Shoals, Alabama, David J. Bruggink and Thomas W.  Toole, Water Quality
and Ecology Branch, Myra P. Smith and  Darrell E.  Houchins, Applied
Research and Education Staff,  TVA's Division of Environmental Planning,
Chattanooga, Tennessee; H.  Brown Wright, TVA's Division of Navigation
Development and Regional Studies, Knoxville, Tennessee; David Stoloff,
Planning and Housing Consultant, Knoxville, Tennessee;  and Edwin B.
Rowland, TVA's Division of Forestry, Fisheries and Wildlife  Development,
Norris, Tennessee.
                               Xll

-------
                             SECTION 1

                           INTRODUCTION
INTEGRATED ASSESSMENT RESEARCH

     The energy crisis has reemphasized the importance of electric power
in maintaining a vigorous economy.  However, because the quality of the
environment must be protected, extensive planning and assessment are
necessary before power generating capabilities can be expanded respon-
sibly.  To meet these objectives, activities involved in system planning
and impact assessment have intensified to the point that significant,
costly delay occurs between the time the need for an electric power gen-
erating facility is first identified and the time the facility actually
becomes operational.  For the typical nuclear plant, environmental analysis
occupies a significant portion of the present lead time of more than ten
years.  Procedures for expediting these activities and for improving their
reliability are essential in assuring adequate electrical power in a cost-
effective, timely manner with adequate assurances of environmental quality.

     When new generating capacity is delayed, the only alternative rapidly
available is the use of combustion turbines.  The cost per delivered
kilowatt-hour for these oil-fired turbines  is more than thirty times that
for hydroelectric power and seven times that for nuclear power at current
fuel and operating costs.

     The Tennessee Valley Authority (TVA) is conducting an Integrated
Assessment Program, with funds administered by the Environmental Protec-
tion Agency (EPA), that is designed to develop methodologies for speeding
the planning and environmental assessment of existing and proposed power
generating facilities.  Integration of activities consist of (1) improving
lines of communications among specialists in planning, engineering design,
and impact assessment, (2) developing a unified data base containing
information for use in a variety of planning and impact assessment activ-
ities, and (3) using more efficient techniques for data display, analysis,
and management decision making.  The work described in this report con-
centrates on the third activity and specifically emphasizes the exploitation
of computer graphics.

EVOLUTION OF ENVIRONMENTAL ANALYSIS THROUGH
THE USE OF COMPUTER TECHNIQUES

     For more than a decade, aeronautical,  civil, mechanical, and architect-
engineers have benefited from analytical applications of computer graphics,

-------
                              -2-
of direct benefit to environmental engineers or scientists have hardly
been exploited.  The intent of this project is to investigate and develop
applications of computer graphics that can contribute directly to reduc-
ing delays in designing and assessing environmental impacts of power
generating systems.

     This work is being performed against the backdrop of practical analy-
sis applications experienced during the planning, design, and operation of
energy systems by TVA, the Nation's largest producer of electric power.
Techniques that have been developed are general enough to be applicable to
organizations other than TVA that are involved with impact analysis of
energy systems.

     An integrated approach to the assessment of environmental impacts of
power generating systems requires the timely consideration of alternative
pollutant controls, dispersal of pollutants in the environment, impacts on
human health and ecosystems, and the attendant costs and benefits.  Many
analyses involved in the assessment process can be conveniently performed
with the use of computational and data management capabilities of the com-
puter.  The full potential of the computer has not been realized because
a sizable delay was and still is involved between the time an analysis is
requested and the time the processed information is presented in suitable
form to the engineer, scientist, or manager.  This delay resulted from
two causes.  First, computer job processing was commonly conducted in a
batch mode:  Computer programs and data were keypunched and then submitted
to the computer for processing; the output was returned to the user at
some later time, completing only a single iteration.  A second cause for
delay was the need for various transformations of the computer output
before an analysis could be accomplished:  Results from computer processing
were commonly presented in a tabular or graphical form selected before the
analysis was begun; if the results of the analysis suggested presentation
of the data in some format (graphic or tabular) other than the one selected,
the analysis was repeated, or the data display was revised manually.

     Time-sharing systems have greatly reduced some of the delay involved
in running many computer programs.  However, interactive graphics terminals
make available even more possibilities in accessing the power of the com-
puter.  Engineering analysis results can be viewed immediately in graph,
picture, or map form on a cathode ray tube (CRT).

     The type of display, its orientation on the screen, and particular
data sets can be selected readily.  A hard copy of the information on the
screen that is suitable for many reports can be obtained at the touch of
a button.  For camera-ready copy, a display can be produced immediately
by means of various interactive digital plotters.  Hardware capabilities
already surpass the available software.  Also, achieving the full benefits
of such capabilities is hampered by the present lack of manpower to imple-
ment imaginative applications.

-------
                                -3-
INTERACTIVE COMPUTER GRAPHICS

     Computer graphics encompasses any visual medium through which man
and computer can communicate.  The computer can be instructed to carry
out tedious calculations and data management tasks, and it can process
results into a meaningful form, such as graphs, maps, or pictures.   A
user can then respond to this information by (1) slightly modifying
specific parameters and rerunning the analysis, (2) changing the type
of analysis entirely, or (3) drawing conclusions and making decisions
based on the results of the analysis.

     Figure 1 depicts the spectrum of currently available devices for
computer graphics output.  True graphics output devices are those that
can be used to draw straight or curved vectors such as graphics terminals
and pen plotters.  Quasi-graphic devices, such as line printers and alpha-
numeric terminals, have the capability of drawing alphanumeric characters
at specific locations.

     Another distinction between devices for graphic output is based on
the computing environment in which they are used; that is, whether the
work is predominantly batch or interactive.  Batch mode is the process by
which a job is submitted to the computing system for processing, and after
relatively long periods of time, the output is returned to the user.  Pen
plotters and line printers are commonly used in this manner.  Output from
the computer is often stored temporarily before being processed into final
form off-line.

     The term "interactive" implies  rapid response to commands that enables
"conversation1' with the computer.  Some output devices, such as graphics
terminals, some electrostatic plotters, and teletypes, can be termed inter-
active.  The working definition of the terms batch and interactive depend
greatly on the experience of the user and the nature of the particular
response requested.  A person who commonly waits a week to obtain a copy
of printed output from a computer job might view a system as interactive
if the job cycle time were reduced to one day.  On the other hand, a person
who is accustomed to computer responses of less than one second to simple
commands from some input device such as a keyboard would begin to question
the interactivity of the system if this response were lengthened to ten
seconds.  If the user feels that a particular task is difficult  (e.g.,
program processing, calculations, display presentation), a longer response
time will be tolerated.  Tasks that  the user views as simple (e.g., com-
puter response to selections from a  menu of commands or editing) must
take a much shorter time to be considered interactive.

     An effective dialog between user and computer depends not only on the
availability of suitable output devices but also on the use of hardware
that facilitates information input to the computer system.  Typical batch-
oriented input devices include punched cards, paper, and magnetic tapes.
Devices that promote an interactive  environment include an alphanumeric
keyboard, CRT, function switches, joystick, light pen, graphics tablet,
digitizer, trackball, and touch panels.  Even direct verbal communication
is possible.

-------
   -

  •8
  n
  n>
  o
tf
H-
O
O O
C H,


T3 O
  n
  p
  •<

  a>
  re

  o.
  81
  M

  O
  O
  f
  n
                                             BATCH
    TRUE

   GRAPHIC
QUASI-GRAPHIC
     FLAT-BED PLOTTERS


COMPUTER ON MICROFILM (COM)
       LINE PRINTER
                                                                INTERACTIVE
    GRAPHICS TERMINALS


  ELECTROSTATIC PLOTTERS
ALPHANUMERIC CRT TERMINALS


         TELETYPE

-------
                              -5-


                             SECTION 2

                            CONCLUSIONS

1.    Computer graphics is an important means for enhancing and
     expediting many types of environmental analyses.

2.    Computer graphics can improve the interaction between the technical
     specialist and the power of the computer.

3.    Computer graphics is an effective medium through which complex
     information can be communicated to managers, planners, and the
     public.

4.    Computer graphics can be applied to techniques for data display,
     interactive analysis, and mathematical modeling.

5.    Computer graphics can result in a savings of analysis time and
     cost.  Analyses previously considered impractical can be readily
     accomplished.

6.    Although some problems can be anticipated in implementing practical
     applications of computer graphics on conventional time-sharing
     computer systems, hardware and software capabilities exist which
     offer effective solutions to these difficulties.

-------
                              -6-


                              SECTION 3

                          SCOPE OF RESEARCH
RESEARCH OBJECTIVES AND GOALS

     Although this investigation encompassed nearly the entire spectrum
of computer graphics devices, initial work emphasized (1) development of
an interactive analysis capability with graphics terminals and various
input devices and (2) exploration of potential applications for
environmental analysis.

     Interactive computer graphics provides the potential for almost
immediate analysis.  Although computing costs will increase as analyses
become more ambitious, savings in the overall cost of analysis can be
achieved.  More sophisticated (and otherwise impractical or impossible)
analysis procedures can be accomplished quickly and inexpensively as
projects move toward completion.  An overall analysis methodology can be
adapted logically and dynamically as intermediate results suggest the
application of other specific procedures or data displays.  Various
design alternatives and system constraints can be evaluated.  The engineer,
scientist, or manager can concentrate on the meaning of an analysis rather
than on manual processing of computer output data or preparation of sketches
for a draftsman.

     During the first two years of work, efforts were directed toward
accomplishing four specific objectives:

1.    To develop methods that would use computer graphics to expedite
     critical environmental analyses of electric power generating
     facilities.

2.    To demonstrate to potential users the range of graphics hardware
     and software techniques that can be practically applied to
     environmental analysis.

3.    To demonstrate the range of environmental analyses to which computer
     graphics can be applied.

4.    To demonstrate the manner in which environmental analysis, through
     computer graphics, can improve the planning, design, and operation
     of energy systems.

     At the onset, it was apparent that successful implementation of
practical demonstrations of computer graphics depends on three factors:

-------
                              -7-
1.   Identification of critical activities in environmental analysis
     that could be markedly improved through computer graphics.

2.   Ready availability of suitable graphics hardware and software
     to potential users as a prelude to the purchase of commercially
     available software.

3.   Identification and correction of a number of problems in computer
     systems.

     To stimulate the identification of potential applications,  seminars
illustrating the capabilities of computer graphics were provided to TVA
employees involved in activities of environmental analysis.  Two video-
tapes illustrating graphics analysis systems were shown to over 300 people
throughout TVA.  In response to a questionnaire requesting suggestions
for valuable applications, TVA's Division of Environmental Planning
identified more than 65 specific potential applications related to
environmental analysis.  These applications are generically summarized
in Appendix A.

SELECTION AND DISTRIBUTION OF COMPUTER HARDWARE AND SOFTWARE

     The events leading to certain configurations of computer hardware
and software are detailed here to provide a guide to others who may
attempt to implement similar graphics capabilities.  During late 1975
and early 1976 commercially available computer graphics terminals were
reviewed.  Although some details are omitted and professionals in the
field will note certain exceptions, the discussion suggests the available
choices among hardware in common use for engineering analysis.  On a
practical level, there are basically three possibilities, each with its
advantages and disadvantages:  the raster scan (video), random beam
(stroke writing), and storage tube systems.

     Raster scan systems have a CRT display similar to that of a
television set.  Vector graphics are formed by activating a phosphor
dot at the appropriate time while an electron beam continuously scans
the screen along lines called rasters.  Generally inexpensive, these
units possess capabilities of color and shading.  However, the raster-
type display may not be as pleasing to the eye as one generated with
continuous lines.  Programming such a system may impose burdens.  Light
pen interaction introduces more complications with raster scan systems
than with other systems.

     In random beam graphics systems, the display is composed of lines
that are redrawn rapidly on the CRT so that the eye sees a constant
display.  As more information is displayed, the rate at which the display
is redrawn (refresh rate) is reduced, resulting in an undesirable flicker-
ing effect.  Random beam systems tend to be expensive ($30,000 or more)
because they require electronics for storing and refreshing a given dis-
play.  Light pen interaction and displays involving dynamic motion can
be readily accomplished.  Software for these systems tend to depend more
on the specific application and therefore are less available for general use.

-------
                               -8-


     Storage tube display systems, which have become quite popular
recently, have (1) the capability of unlimited display without flicker,
(2) low cost, (3) excellent resolution, and (4) compatibility with
numerous software packages.  However, storage tube systems sacrifice
brightness (contrast) and offer no shading capability, color, or
selective screen erase.

     The graphics capabilities needed initially to accomplish the
research objectives were reviewed.  Capabilities for color, shading,
animation, and light pen interaction were not judged to be necessary
for many initial demonstrations.  We also wished to make graphic terminals
available to a large number of different potential users, partly through
project funds and partly through the responsiveness of other TVA programs
to the needs expressed by users.  As a result of these factors, storage
tube displays were selected for initial graphics implementation.  The
terminal selected was a Tektronix 4014-1 featuring a 19-inch (diagonal)
screen, four character sizes, upper and lower case letters, and 780 x 1024
addressable p'oints (enhanced mode 3120 x 4096).1  Programming of this equip-
ment is similar to other types of graphics software (e.g., CALCOMP) with
which many users are already familiar.  A complete graphics station,
including a flexible-disc memory unit, graphics tablet (digitizer), and
hard-copy device (digital plotter), was obtained for about $25,000.

     Several graphics software packages were available that would make
possible the demonstration of a broad range of graphics capabilities
when implemented on TVA's IBM 370/165 computer system with the time-
sharing option (TSO).

     As a result of our inventory of applications, direct assistance was
provided to establish five graphics hardware stations in TVA.  Table 1
shows the location of each station, the distribution of equipment,
and funding arrangements.  In several cases, project funds were used to
purchase equipment to supplement existing graphic hardware purchased by
other means.  In one case, project funds were used to purchase the entire
graphics station.  In another case, no project funds were expended, but
assistance was provided in preparing the necessary justification for
obtaining the equipment.  New graphics software was purchased by TVA's
Computer Services Branch from other TVA funds.

     Figure 2 shows relationships among the various graphic software
packages that have been used to develop demonstrations.  All routines
are callable by the user in FORTRAN IV.  Tektronix Plot-10 Terminal
Control System (TCS) contains the basic graphics software, including
routines for windowing, clipping, moving the graphics beam, drawing
vectors, and initiating input and output operations.  A user's appli-
cation program whose output is to be viewed on the CRT is processed
through this group of routines.  The Plot-10 Advanced Graphing Package,
AG-II, allows ready generation of conventional line, point, and bar
graphs that can be tailored to a particular need.

-------
                                                                           Computer  equipment9
       Graphics  station  location
           and organization
Water Quality and Ecology Branch
Division of Environmental Planning
401 Building
Chattanooga, Tennessee
Air Quality Branch
Radiological Hygiene Branch
Division of Environmental Planning
River Oaks Building, Muscle Shoals
Alabama
Regional Studies Branch
Division of iJavigational Development
Regional Studies, Liberty Ruilding
Knoxville, Tennessee
Forest and Wildlife Resources Branch
Division of Forestries, Fisheries
and i-Jildlife Development
.Jorris , Tennessee
'.Jater Quality and Ecology Branch
Division of Environmental Planning
E&D Building
Muscle Shoals, Alabama
0
0
X
0
0
X

X



X

X

X





X

X

0
X
V
/*
0
0
0




X





aSymbols:   X--purchased or leased  with  the  use  of  research project  funds; 0--obtained by other funding arrangements,

-------
                              -10-
       CALCOMP
       GENERAL-
       PURPOSE
       PACKAGE
       CALCOMP
         3-D
     PERSPECTIVE
       DRAWING
       SYSTEM
      TEKTRONIX
       PLOTTER
       UTILITY
      ROUTINES
      TEKTRONIX
     DIGITIZING
      ROUTINES
      TEKTRONIX
      FLEXIBLE
        DISC
      ROUTINES
   USER-DEVELOPED
       GENERAL-
       PURPOSE
       GRAPHICS
       ROUTINES
CALCOMP
UTILITY
PLOTTING
ROUTINES


TEKTRONIX
CALCOMP
PREVIEW
ROUTINES
                                                I
  TEKTRONIX PLOT-10

      TERMINAL

       CONTROL

    SYSTEM (TCS)
                       TEKTRO

                          CRT

                        OUTPU

                        DISPL
TEKTRONIX
ADVANCED
GRAPHICS
  AG-II
   USER-
  WRITTEN
APPLICATION
 PROGRAMS
Figure 2.  Relationship between TVA's graphic software packages.

-------
                              -11-
     The Plot-10 Graphics Tablet (Digitizer) and Plot-10 Flexible Disc
Utility Routines allow the user to communicate readily with this hardware
and exercise its various operational modes.  Tektronix Plot-10 CALCOMP
Preview Routine permits programs written with CALCOMP's utility plotting
routines2 to be viewed on a Tektronix graphics terminal with few, if any,
program changes.  TVA has available two other CALCOMP graphics software
products, a general-purpose contouring package (GPCP), and 3-D perspective
drawing software.

-------
                               -12-


                             SECTION 4

              DEVELOPMENT OF GENERAL-PURPOSE GRAPHICS

     One group of computer programs that operates in this software
environment deserves particular mention.  Some of these routines were
developed to make graphics easier to use.  Other routines were formu-
lated to preclude the development of similar graphics routines by indi-
vidual users.  These routines and their capabilities are described
below.

     The computer-generated displays presented throughout this report
represent either a copy of the CRT display shown to the user or a com-
posite made up of several CRT displays and combined manually.  In most
cases annotation for the figures was added by manual methods (computer-
generated text cut and pasted into position).  Programming the annotation
required for publication by using the currently available low-level soft-
ware is a rather tedious and time-consuming activity.  This fact repre-
sents a prime limitation for providing "complete," computer-generated,
camera-ready report figures.  However, more sophisticated graphics
software to facilitate figure annotation is beginning to be offered
commercially.  The use of these capabilities will be explored in the
near future.

INTERACTIVE GRAPHICS DISPLAY PACKAGE (IGDP)

     Although computer graphics can often shorten the time required for
analysis, the overall cost for the analysis may not be reduced.  Although
requirements for manpower and elapsed time are reduced, computer processing
cost may increase.  To keep computer time charges at the lowest possible
level, an interactive graphics software package that is simple, inexpensive,
and easy to use was developed to display three-dimensional data (i.e., one
variable as a function of two others) with (1) a minimum of program develop-
ment time,  (2) a minimum amount of data manipulation to generate the
required input format, and (3) a maximum degree of flexibility for genera-
ting a display of the required complexity to convey technical meaning and
thus minimize the amount of computer processing required.  This work was
deemed necessary to stimulate interest in applications using this display
capability before the purchase of more flexible, but expensive, graphics
software was justified.

     The IGDP is made up of five basic graphics routines:  (1) numeric
data presentation routine; (2) partitioned contour plotting routine;
(3) conventional line contour routine; (4) three-dimensional surface
plotting routine; and (5) two-dimensional sectioning routine.

     Overall control of the programs and file allocations are accomplished
for the user automatically through a command language program (CLIST).
Selection of any or all of the various graphics modules is made by
answering yes-and-no questions.  Instructions explain the various graphics
options available within a particular routine.

-------
                               -13-
     Input to this package consists of a regular,  rectangular (X,Y)
array of data corresponding to vertical coordinate (Z)  values.   Array
sizes (X,Y) up to 250 x 250 have been used.   Interpolation of values  for
missing data and selection of regular increments for X and Y data are
intentionally left to the user because the particular technique for
interpolation depends on the particular application.  Including routines
for interpolation and incrementation in the graphics package would have
complicated the problem (e.g., delayed the provision of some early
capabilities and increased programming complexity) and added little  to
the graphics output.

     The user may also alter his input data format to illustrate a
particular feature of the data more effectively.  This is particularly
true for three-dimensional displays.  For example, the user may find
that, by surrounding the array of data with some meaningful reference
value, he can make relationships of various points of the generated
surface clearer as compared with those on an arbitrary surface in space
with no points of reference.

     Use of the specific graphics routines is illustrated below.  As an
example, a data set was created that contained the concentration of a
chemical dye measured at a particular time in the surface water near a
series of waste discharges and water intakes along a reach of the Tennessee
River.

1.   The Numeric Data Presentation Routine is used to display the complete
     array of input data in an orderly format selected by the user.   The
     user can also store this reformatted array as a data set on the
     flexible-disc memory unit or as a member of his TSO library.
     Figure 3 shows an annotated copy of the CRT display of input data
     for the example.  The arrangement shown can be considered as a plan
     view of the data.

2.   After the Partitioned Contour Plotting Routine calculates maximum
     and minimum values of the data, it prompts the user for the  desired
     number of intervals into which this range of data is to be divided.
     The data values are then partitioned into these intervals, and a
     letter or symbol can be plotted at each corresponding data location.
     A key for interpreting the plot is also given.  Figure 4 shows the
     sample data presented in this form; ten intervals were selected.

3.   After the Line Contouring Routine calculates minimum and maximum
     values of the data, it prompts the user for the number and values
     of the contour intervals to be plotted.  For a rectangle composed
     of data values at each corner, a simple calculation determines
     whether each of the specified contours crosses through or is con-
     tained inside the figure.  Intersections of a  contour with the
     rectangle's boundaries are then connected by straight lines.
     Figure 5 illustrates this procedure for the sample data.

-------
A
W
H H
D Z
O¥UJ

IiKliJ
MM ^% VU
M/Dft?
Z 0
fm \J
^¥7
^% *• fc
l_ f\ LJ
wff
WK
i i a, |
HU.I-
*\ •
u u.

<9
RIVERBANK C0.1ml. INCREMENTS^

9.9
9.9
0.5
e.s
1.9
2.e
2.3
2.6
2.9
3.9
3.8
3.2
3.5
3.5
3.5

9.9
9.9
9.6
1.6
2.7
2.7
2.8
2.9
3.1
3.3
3.5
3.6
3.8
3.8
3.8

9.9 e.e
9.9 9.9
9.9 7.9
1.9 6.8
2.2 6.4
2.5 9.0
2.7 11.6
2.9 14.5
3.5 14.4
4.1 14.3
4.9 14.2
5.4 7.1
5.8 9.9
5.2 e.0
4.6 e.0

e.e
e.e
1.8
1.0
9.5
e.e
e.e
e.e
e.e
e.e
e.e
e.e
e.e
e.e
0.0

e.e
e.e
2.6
e.s
e.7
e.e
e.e
0.0
e.e
e.e
e.e
e.0
e.e
e.e
e.e

e.e
e.e
11.0
11.0
1.2
e.s
0.5
0.2
e.e
0.0
e.e
e.e
e.e
e.e
e.e

e.e
e.e
31. e
29.2
16.4
0.6
0.2
0.2
0.1
e.e
e.e
9.9
9.9
9.9
9.9

9.9
9.9
25.0
19.0
14.2
9.9
5.6
1.3
1.0
0.2
0.0
0.0
0.0
9.9
0.0

0.0
e.0
18.0
3.6
0.8
0.2
0.2
0.1
0.1
0.1
0.1
9.9
9.9
e.0
0.0

10. e
e.e
49.0
95.0
2.4
0.2
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0

0.0
e.e
e.e
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0

0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
9.9
9.9
0.0
0.0
0.0

0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
9.0
0.0
0.0
0.0
0.0

0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
                                     CONCENTRATION Cmo/1
Figure 3.   Formatted tabulation  of  concentrations of dye in surface water,  as  generated by
           IGDP software.

-------
                       RIVERBANK CO.I  Mll«3

   AAA-AAAAAABAAAA
   AAAAAAAAAAAAAA
z
Q
y

\-
-------
                                    -16-
                                                X .*Si!:.:i--:-:-:-:^--££ .'iv.
                                                ••  ,«tiace

-------
                               -17-
4.   The 3-D Perspective Routine plots the data viewed as  a  continuous
     surface.  The user then has numerous options for manipulating the
     display, including figure rotation, surface tilt, distance of eye
     from figure, position and size of figure on screen,  exaggeration of
     Z axis (expansion around mean data value), automatic or manual
     scaling of data by the drawing routine,  and degree to which the X,Y
     directions are scaled with respect to the Z data values.  Figure 6
     shows a three-dimensional view of the sample data.

5.   The 2-D Cross-Section Routine allows the user to interactively
     obtain cross sections of the surface orthogonal to either the
     X or Y axis.  Profiles may be overlaid on each other for com-
     parison.  Figure 7 shows cross sections at several stations
     perpendicular to the river bank.

KIVIAT DIAGRAM PLOTTING ROUTINE

     Many scientific and engineering problems involve the study of
complex systems.  Analysis of systems impacted by man is now receiving
considerable attention.  The objective of a typical analysis is to
determine and evaluate changes in the states of the systems with respect
to time and space.  To accomplish this task, many variables must be
considered simultaneously.  For example, to access changes in the eutrophic
conditions of a body of water, the physical, chemical, and biological
characteristics are used to describe the state of the system.  To assess
the impacts of a power generating facility on local socioeconomic conditions,
the state of that system might be represented by variables quantifying
the availability of facilities for education, health care, recreation,
transportation, housing, and community services such as garbage collection,
police and fire protection, water supply, and sewage disposal.  Social
attitude and political and sociopsychological factors may also be important
in some analyses.

     Many statistical and mathematical techniques have been used to
quantify and visualize sets of multivariate data.3  Diversity indices
and clustering techniques are numerical methods commonly used by biolo-
gists.  Various graphics displays have also been developed.  Anderson4
has suggested the use of circular glyphs to represent multivariate data.
Chernoff and Rizvi5 have described the formulation of a cartoon face for
simultaneously displaying as many as 18 variables; the quality of the
resulting face communicates information about a system.

     Morris6'7 has described the use of a graphical configuration (Kiviat
diagram) in which variables are scaled and plotted along axes oriented
as spokes of a wheel; when the data points are connected, a characteristic
polygon results.  A similar technique has been applied to medical research,8'9
to aquatic chemistry,10 and in the presentation of sensitivity analysis
results generated with an aquatic systems model.11  In the latter case,
365 axes, representing days of the year, were distributed in a circular
configuration.

     At the top of Figure 8 is shown one form of the Kiviat diagram
adapted to socioeconomic impact analysis.  Six variable axes, each
representing a different indicator of socioeconomic conditions, and two

-------
  18ft
                                                                                                          00
Figure 6.  Three-dimensional display of concentrations of dye in surface water, as generated

-------
   o
   E
     M
   o
   H
   UJ
   O
   KEY

STATION 11
STATION  h'
                                                                                           VD
                        DISTANCE OUT FROM RIVERBANK
                        IN  Y DIRECTION  UNITS C10 ft/unit)
Figure 7.  User-selected profiles of dye concentrations perpendicular to the river bank, as
         generated by IGDP software.

-------
                               -20-
vertical dummy axes were used.  The six county-wide variables selected
here might be used to measure the desirability for establishing family
residences, which may, in turn, be important to an industry or government
considering relocation or expansion of activities.  The six indicators
selected are (1) average teacher's salary, (2) average family income,
(3) police expenditure per capita, (4) unemployment rate, (5) percent on
welfare, and (6) housing vacancy rate.  Another group of variables
tailored to siting situations for other facilities could have been used
just as easily.

     Inspection of the selected indicators in this sample case revealed
that the display could benefit from logical grouping into two categories.
The first three indicators might be termed "desirable" county features,
whereas the last three indicators might be termed "undesirable."  County
ratings for these two categories of variables were grouped on the left
and right sides of the diagram.  On the left side, indicators (1) through
(3) were plotted so that a rank of one (highest-ranked county) was
plotted along the particular axis at a distance equal to the radius.  A
county with average rank was plotted at a distance equal to one-half the
radius.  The county with the lowest ranking was plotted at the circle's
center.

     Indicators (4) through (6) were plotted differently, since high
rank indicates a relatively undesirable condition; for these the plot-
ting scheme is inverted.  The county with the lowest percentage of its
population on welfare, for example, was plotted at a distance equal to
the radius.  The county with the highest welfare rate was plotted at the
center of the circle.  Thus, when all the points are connected, including
points arbitrarily near the two central vertical axes, a closed polygon
results.  The resulting figure resembles a butterfly for a county that
rates high in all categories.   Stated another way, the better the county's
overall ratings, the more the figure will achieve a well-developed
butterfly shape.  Other characteristic images can be constructed by
reclassifying and rearranging variables and by using other plotting
schemes.

     At the bottom of Figure 8, the results of the analysis for eight
actual counties (ranked among 22 East Tennessee counties) are shown.
Numerical results of this are shown in Table 2.  Note that the data are
more difficult to interpret from the numeric listing than from the
graphic display.  It is readily apparent from the figures that counties A
and B are more desirable places to live than are counties E and H.
County C rates very low on the "desirable" indicators and average on the
"undesirable" indicators.  County D is average with respect to both sets
of indicators.  The utility realized by this graphical technique depends
on the user's ability to develop an arrangement of the display that
conveys meaning.  Flexibility can be achieved by varying the number and
type of variables, spatial relationship of the axes, and the location
and number of dummy axes.  Actual scaled numerical values, variances, or
residuals can be plotted instead of rank.  Other figures, ranging from
arbitrary polygons to highly structured figures such as multipointed
stars, can be formed.

-------
     COUNTY A
COUNTY B
                                                             COUNTY C
                                                                                         COUNTY D
                       AVERAOE TEACHER'S
                          SALARY
               MEDIAN FAMILY INCOME
                        f. I 9785
               f •xp«n«*l
                       POLICE
                                                              JNCMPLOYMCNT RATS
                                  POPULATION RECEIVING
                                      WELFARE
                                                              HOUSING VACANCY RATE
                                                                                                      I
                                                                                                      Si
   t-;^;
      COUNTY E
                                 COUNTY F
                           COUNTY G
                                                                                        COUNTY H
Figure 8.   Buttetfly configuration of a Kiviat diagram applied to the description of a
           socioeconomic system.

-------
TABLE 2.  NUMERIC RANKINGS AMONG 22 COUNTIES FOR SIX SOCIOECONOMIC INDICATORS
Name
A
B
C

D
E
F
G
H
Percent on
welfare
7
3
11

4
5
21
8
1
Percent
unemployment rate
10
22
11

12
15
4
6
1
Average
teacher's income
2
1
21

12
18
13
16
20
Average
personal income
2
1
19

16
20
14
18
22
Police
expenditures
per person
2
4
22

13
21
14
7
17
House vacancy
rate
8
4
14 N
i
15
2
9
22
1

-------
                               -23-
     Figure 9 shows an example of a star shape that can be used to
illustrate the condition of an education system.   Information about
the system can be inferred from the degree to which the resulting
polygon has achieved a given shape and from the symmetry,  area, or
centroid location of each figure.  A series of figures can be used
to represent progression of the system in time or space.

     To explore the use of various types of diagrams, an interactive
computer graphics routine was developed.  The user conversationally
enters the necessary display information at a computer terminal.  The
resulting display is shown immediately on a graphics CRT.   An experimental
plotting variation can then be studied.  The user can also interface
this display routine with the results of numeric computation routines,
allowing complex data manipulation to drive each new display.

     Although the examples discussed above used socioeconomic data,
other types of data common to environmental and management analysis
have been displayed by means of the Kiviat diagram programming routine:

1.   The variance of monthly samples of benthic insects at six sample
     locations (one figure per month).

2.   Chlorophyll composition in field water samples  (one figure per
     depth).

3.   The measured concentrations of sulfur dioxide leaking around the
     circumference of a stack (one figure per elevation level).

4.   Economic, environmental, and engineering factors relevant to
     implementing air pollution control alternatives  (one figure per
     generating facility).

5.   Five metallic elements obtained from core drillings  (one  figure
     per sample depth).

6.   Nine physical and chemical variables of water quality (one  figure
     per sampling station and time).

7.   The variation in the predictions of a water quality model after
     a small perturbation in input (one figure per year's simulation).

8.   Managerial skills of project personnel (one figure per person).

9.   Elements of costs at a given coal-fired steam plant.

One sample of each display is shown in Figure 10.

     Experience has shown this technique to be more  useful for some
applications than for others.  Usefulness depends on the nature  of
the data itself, the level of efforts applied to develop a suitable
display scheme, and judgment in the selection of variables.  Nevertheless,
this simple graphical technique can lead to new and  valuable insights  into
the meaning of a given set of data.

-------
        COUNTY A
                            AVERAGE
                         TEACHER  SALARY
                                           COUNTY B
                                                       PERCENT STUDENTS
                                                         OVERCROWDED
                                                                              COUNTY C
EXPENDITURE PER
     PUPIL
                                                                                                                  COUNTY D
                                                        PUPILS PER TEACHER
       COUNTY E
                                          COUNTY p
                                                                              COUNTY G
                                                                                                                  COUNTY H
Figure 9.   Star  configuration of a Kiviat  diagram depicting county-level  educational indicators.

-------
         MONTH AT SIX  RIVER MTLT
         STATIONS—OCTOBER 1973
                                          OF A WATER SAMPLE BY
                                          DEPTH—1  m
                                                                    "HDUJTCCITT iu
                                                                    BY ELEVATION—50  ft
       RH 241.7
       RM 240.9
   RM 242.5
   RM 240.71	/—¥—:>--RM 243.8
   RM 239.0
           CHLORO a
           CHLORO b
                                                          } CHLORO c
(4)  DECISION  FACTORS FOR IMPLEMENTING
    AIR POLLUTION CONTROL BY PLANT--
    KINGSTON  STEAM PLANT
     POLITICAL
   CONSTANT
SOg CONTROL
    COSTS
        ANNUAL
     SDEL COST
   STACK HEIGHT
       OPULATION
   AMBIENT S02
                                    (5)  METAL CONCENTRATIONS IN
                                        WELL SAMPLE CORE BY DEPTH-
                                        5 m
                                                               Cr
                                                             (6)  NSF WATER QUALITY
                                                                 INDEX FACTORS BY
                                                                 SAMPLE—WATER SAMPLE III
  (7)  RATIO OF STANDARD SIMULATION
      TEMPERATURE TO SIMULATION WITH
      5° INCREASE IN UN FLOW WATER
      TEMPERATURE BY ELEVATION—25  m
       JUL
I--JJAN
                 OCT
                                   (8) MANAGEMENT LEVEL CAPABILITIES
                                      BY EMPLOYEE—EMPLOYEE A
   DELEGATION OF
  RESPONSIBILITY.

DEVELOPMENT OF
  SUBORDINATES
                                                   COMMUMI CATION
DECISION
MAKING

 ACCOMPLISHMENT
 OF OBJECTIVES
                                                         USE OF RESOURCES
                                                           (9) MONTHLY STEAM PLANT COSTS
                                                               EXCLUDING FUEL BY DATE AND
                                                               PLANT—JUNE  1977, COLBERT
                                                               STEAM PLANT
   LABOR
SUPERVISION
  & ENGR.

OPERATING


MAINTENANCE
              Figure  10.   Nine examples of  the  use  of  the Kiviat  diagram  for  engineering and management
                             analyses.
 SUPPLIES
SUPERVISION
  & ENGR.

  OPERATING
                                                                                                                                 MAINTENANCE

-------
                               -26-
INTERACTIVE AG-II PLOTTING ROUTINE

     A general-purpose program for plotting data was developed that allows
the user to interactively generate a conventional plot of data that is
tailored to a particular need.  The program is intended for use by
individuals familiar with TSO and plotting options with Tektronix Plot-10,
AG-II Software.  Figure 11 shows two plots of three sets of data displayed
first as a bar chart and then as a line plot.  Data to be plotted can be
input to the program from either the terminal or a data set in the user's
library.  Data can be specified as X,Y pairs or as X and Y arrays.  Several
sets of data can be read to create overlays of different data.  Fourteen
basic options are avaiable to alter and label the figure (Table 3).
This program has found wide application because it provides for rapid
viewing of data, eliminating the need for developing special programs.
TABLE 3.  USER-DEFINED OPTIONS FOR GENERAL-PURPOSE PLOTTING PACKAGE
          Parameter
Option
Character size for labeling plot
Size (type) of major tick marks on X and Y axes
Size (type) of minor tick marks on X and Y axes
Type of scale along X and Y axes (linear, log, etc.)
Suppression of zero value labeling
Type of symbol plotted for each data value
  (including bar chart shading)
Size of data symbol
Type of connection line through data
  (including bar chart selection)

Missing data value designators
Draw frame around plot
Width of individual bars
Labels for X and Y axes
Scaling minimum and maximum X and Y axes

Screen minimum and maximum coordinates in X
  and Y directions
4 sizes
6 types
6 types
8 types
  levels
   types
2
11
(15 bar shades)
Continuous
9 levels plus any
  user-defined
  line type
Variable
2 levels
Continuous
20-character maximum
Calculated from data
  or user-defined
0-1024 (X)
0-780 (Y)

-------
   SB
VI

H
Z


LU
h-
   28
   ie
                                   BAR CHART
                                                             40
            W

            H
            UJ

            <

            H  28
                                                             19  -
                                                                                              LINE PLOT
                      10       15

                 ABSCISSA  UNITS
                                      20
25
                                  18       15

                              ABSCISSA  UNITS
                                                                                                                   I
                                                                                                                   to
   Figure 11.   Sample plots of data using the interactive AG-II  plotting routine.

-------
                               -28-
DIFFICULTIES INVOLVED IN IMPLEMENTING GRAPHICS HARDWARE AND
SOFTWARE

     At the onset of this project, several difficulties were encountered
that could potentially inhibit practical implementation of certain
demonstrations.  Later, some difficulties arose as the result of the
demands placed on TVA's computer system by rapid increase in use of
interactive computing.  Other difficulties resulted from problems in
the Tektronix software as implemented on TVA's computer system.  Finally,
as user interest intensified, there were problems in meeting the specific
demands of certain applications by means of the existing combinations of
hardware and software.  Appendix B lists these difficulties and describes
approaches to solutions.

-------
                               -29-


                             SECTION 5

       APPLICATIONS AND DEMONSTRATIONS OF COMPUTER GRAPHICS
     This section details representative types of applications that have
been developed for various aspects of environmental analysis of energy
systems.  Data display, interactive analysis, and interactive modeling
are discussed.

     It should be reemphasized that the graphic displays shown in Section 5
represent either a copy of the CRT display shown to the user or a composite
made up of several CRT displays and combined manually.   In most cases,
annotation for the figures was added after the graphics were produced to
meet specific requirements for publication.  Text was computer-generated
and then affixed to the basic figure in an appropriate location.

AIR QUALITY ANALYSIS

Display of Data

     Many types of data are used to analyze air quality.  Figure 12
shows a display of measured ground elevation and, below it, the measured
depth to bedrock.  The data were gathered by actual field measurements
to investigate the reason for wide variations in growth of soybeans on
adjacent test plots that received identical care.  These plots formed
the test area to (1) study the effects of sulfur dioxide from coal-fired
steam plants on crops and (2) determine the effectiveness of a local air
cleaning system for controlling crop exposures.12  Using the three-
dimensional data displays made soil depth in the vicinity of the experi-
mental plots easy to visualize.  Those test plots with shallow soil
depths produced less soybean growth than those with greater soil depth,
as might be expected.  This analysis was useful for designing future
experiments.

     A second application arose for the same research project on air
quality effects--a study of dose-response kinetics through controlled
concentrations of sulfur dioxide, nitrite, nitrate, and ozone in a
vegetation growth chamber.  To "calibrate" the chamber, representative
environmental variables--illumination, relative humidity, temperature,
and ozone concentrations—were measured at several locations in the
enclosure.  Figure 13 shows a sample of the illumination data.  The
figure clearly illustrates that illumination is not uniform, but slightly
higher at one end of the chamber than at the other and slightly higher
along the centerline than along the sides.  This graphic display can
guide readjustments of the lighting arrangement and suggest the magnitude
of this source of variation (illumination) in the statistical treatment
of data.

     Many analyses of air quality require the simultaneous consideration
of different variables measured in the field.  Computer graphics provides
the means for rapidly combining this data in a format suitable for analysis

-------
                           -30-
                                                SURFACE CONTOUR
                                                   BEDROCK
                                                    CONTOUR
Figure 12.   Profile of measured ground elevation and bedrock in the
           vicinity of soybean test plots.

-------
INTENSITY
RANGE>
 180-87
INTENSITY
RANGEi
94 - 73
             Figure  13.  Measured illumination levels at two elevations in a vegetation growth chamber.

-------
                                -32-
Figure  14  shows  a  complex plot for measured data at a particular time at
TVA's coal-fired Johnsonville Steam Plant.  At the top is a bar chart
that relates which of  the plant's ten generator units are operating and
at what particular capacity.  The main portion of the plot consists of
various variables, including potential temperature (a thermodynamic
temperature parameter), epsilon  (a measure of turbulence), wind speed,
and wind direction, plotted against height above ground.  Also shown in
Figure  14  is a plot of the visible plume boundaries and the plume's
centerline.  Such  display routines have greatly reduced the time needed
to present the air quality engineer with field data and to prepare a
report  after completion of the analysis.

Interactive Analysis of the Effects of Mobile Point Sources

     Computer graphics can improve the decision-making process by pro-
viding  a means for establishing a rapid, meaningful interaction between
the scientist or engineer conducting an environmental analysis and the
power of the computer.  That is, interactive computer graphics allows
technical expertise to focus on the meaning of a particular analysis or
on an evaluation of different alternatives while the computer generates
the necessary computations and manages the data and output display.
Figures 15 and 16 show two steps in analyzing distributions of air
pollutants from mobile point sources.   The situation simulated in the
analysis might be encountered during the construction of a power-
generating facility.   Various types of equipment (vehicles, machinery)
that generate gaseous pollutants can be distributed over the construction
site in a given pattern.   A typical analysis entails several steps:

1.   The user is presented with a grid (100 m on a side) on the CRT. If
     needed,  a digitized map of the construction area can be overlaid on
     the grid as a reference.

2.   The user then dynamically positions any number of sources on the
     grid with the terminal's cross-hair cursor.  The characteristics of
     each source (strength, height) and the meteorological conditions
     are then entered from the keyboard.  Figure 15 shows a completed
     source grid.

3.   The computer  then calculates the pollutant concentration at each
     point on a downwind receptor grid (1600 m on a side).  Output from
     the analysis  is available in tabular, contour, or three-dimensional
     format using  IGDP.  Figure 16 shows a three-dimensional representa-
     tion of resultant distributions of pollutants for specified conditions
     of Figure 15.

4.   The user then can repeat the initial step, change the input
     variables,  and again view the output.  Patterns of pollutant
     dispersion  from this analysis can be linked to mapping routines,
     which could overlay patterns of land use, land cover, or
     environmentally sensitive areas.   Models that predict receptor
     impacts could be invoked, and the results displayed.

-------
                                      PLUME RISE DATA
                                    4/24/75-1980 HOURS

(tap.
	 7 	
l
t
1
\
1
" 1
t
\
£ -i
CD
H
UJ
*y~

'
(
2
1
\

	 IBM 1 	
PERCENT-i ^
i [/•
LOAD 1 k
_" g
!
/
/
'
'
•
P
\
2 4
GENERAT

) 1 X
• / : ' ^

tT*ll
H^-AJ
y
.*.-
u
SI
POTENTIAL T
Ja
*-'"

1 1 1 1
e ic
EMPERATURE
^
_,*•'
. *^*~

1 1 1 1
18 IE
CDE6 C) 	
/>! £
\ 0 [: 7 1
/ /> %. \
6 8 IB ' 12 /
tON UNIT ' i / ';
i /
' X '
, 7
, r\j I
1 \ '
; i\. i
-. \ i
-*
	 0
DISTANCE D
1 1 1
ee
1 1 1 1 1 1 1 1 1 M 1 1 1 I 1 1 1 1 1 1 1 1 1 I
» 28 38 32 34 30
EPSXLON«M|/3 	
I""!
""Ji""^''^"",!
)

28
i

3WNWINDCH)
1 1 1 1
88 2Si
1 1 1 1
5
U
n
/\
' /
1 1 1 1 1 1 1 1 1 1 1 1
ee 3888 ssee 4aae
WIND SPEED <«XS> 	
MM 1 M 1 1 1 I 1 I
18 IS 28
2ND DIRECTION CDE6J 	
1 1 1 1 1 1 1 1 1 1 I 1 1 1 1 1 1 1 1 1 1
e ak SB it IBB
                                                                                                      Co
Figure 14.   Display of air quality data in the vicinity of  a  steam plant.

-------
AFFLUENT SOURCE LOCATIONS, POLLUTANT BEING EMITTED,  AND METEOROLOGICAL ASSUMPTIONS

                                                                     SYMBOL

                                                                       o

                                                                      X
                                                                      A
SOURCE STRENGTH WIND SPEED STACK HEIGHT STABILITY CLASS
CO/MC) Cm/me)  CA-F5
1 1090 1 IQ K
2 1000 1 IB F
3 1000 1 10 K
/
c
c
1
*
E
D
3












V
/

(
\


L


f
\

\
f


k









































































                                                                                               I
                                                                                               LO

-------
Figure 16.  Three-dimensional display of downwind pollutant distribution pattern for a
            linear array of three gaseous emissions sources.

-------
                                -36-


     Previously,  this analysis  procedure was conducted with the use of a
hand calculator;  only a  limited number of source configurations could be
tested and manually plotted in  a reasonable amount of time.  Interactive
graphics has made possible more rapid and accurate identification of
potentially adverse situations  and has provided analytical results on a
more timely basis for court hearings involving potential impacts during
construction phases.

Analysis of Output from  the Drift, Vapor, and Dry Plume Model

     To study the environmental impact of heat released from power
plants, models simulating the behavior of buoyant plumes from cooling
towers and combustion stacks are used.  Effluent species include gases,
water vapor, and  water droplets containing soluble and insoluble material
and particulate drift, which remains after evaporation of the water.
The model simulates dynamic phenomena that describe plume trajectory,
dilution, concentration, and receptor deposition as functions of eleva-
tion and distance from the source, source design parameters, and atmos-
pheric conditions.  Such models are adaptable to specific sources such
as dry stack-gas  plumes; a wet  scrubbed stack-gas plume; and vapor and
drift plumes from mechanical- and natural-draft cooling towers.  For
example, to estimate the ground-level deposition of salt or other sus-
pended or dissolved species from cooling tower drift plumes, one must
recognize that the time scale of interest, and therefore the meteoro-
logical data sets, differs from that for the prediction of length of the
visible plume.   The latter is commonly considered, if for no other
reason, because of aesthetic impact.

     Two vapor plume models (developed by Envirodyne Limited, Waterloo,
Ontario) were supplied to the TVA's Division of Environmental Planning,
Air Quality Branch.   These vapor plume models are applicable to mechanical-
and natural-draft cooling towers and to dry and scrubbed stack-gas
plumes.

     An extension of this application was the development of a simple
model for drift plumes that may be applied to cooling tower plumes and
scrubbed stack-gas plumes, provided sufficient information on the source
of the effluent is available.  To economize on computer costs and to
approach interactive iteration, the model was made as uncomplicated as
possible while retaining the necessary and important physics and impact
dynamics of drift plume behavior.  The principal output of the model is
in the form of drift deposition rate (mass per unit area and time) as a
function of radial distance from the source.  However, one may obtain
estimates of excess humidity, temperature, visible plume length, and
other variables,  if needed.  The graphics interface with these models is
described in the  following section.  Specific details of the simulation
models and associated computer  documentation are provided in separate
reports.13'14

     For the vapor or drift plume model, a simulation run can be accom-
plished in about  1 to 2 seconds of computation (CPU) time on TVA's
computer system.  This allows rapid interaction between user and model.
For purposes of this demonstration, an interactive routine for input and
output display was developed for the plume behavior model.  The user is

-------
                               -37-
prompted for wind speed, temperature, relative humidity,  and stability
class.  Other variables (cooling tower design information)  have specified
default values that the user may change.  The model then simulates  the
behavior of the plume and presents the results (Figure 17).   Downwind
plume behavior is shown for both the near-field (0 to 300 m) and the
far-field (300 to 10,000 m) because plume behavior over these two
approximate ranges are modeled differently.  The addition to the rela-
tive humidity at ground level is also shown as a function of distance.
This indicates the possibility of predicting local fogging for the
meteorological conditions specified.  The user can easily make repeated
runs using other conditions.

     To obtain yearly average conditions around the source,  the joint
frequency distribution for the four input meteorological variables  must
be determined.  For example, when the drift model is used,  a total  of
1024 simulations must be run to account for the different probabilities
of occurrence of combinations of four classes of wind speed, four classes
of temperature, four classes of relative humidity, and sixteen classes
of wind direction.  A series of computer programs was developed by TVA's
Division of Environmental Planning, Air Quality Branch, to determine a
representative value for each class variable and to estimate the proba-
bility of occurrence based on observed weather data.  The drift model
was then run for all 1024 cases.  Because of the amount of computation,
these calculations were not run interactively.  The downwind deposition
rates were summed with those of previous runs for each of the sixteen
wind directions.

     An interface routine was developed to format these data for input
to the CALCOMP General-Purpose Contouring Package.  This graphics package
fits a surface to the given data, calculates a regular, rectangular
grid, and plots contours.  Manual techniques, although somewhat tedious,
were also explored for generating a  regular grid.  The resulting grid
serves as input into a previously described Interactive Graphics Display
Package (IGDP) with three-dimensional plotting capabilities.  Figure  18
shows a contour plot of a sample drift pattern; Figure 19 shows the same
data plotted in three dimensions; and Figure 20 shows deposition rate
along several sectors.

     Although output data from this  simulation can be manipulated inter-
actively, a direct, interactive interface with the model itself would be
highly desirable for meeting anticipated analysis requirements.  A
sensitivity analysis to evaluate the influence of terrain, number and
type of sources, and source arrangement at the site could be accomplished
more rapidly.  Also, interactive model-to-graphics capability would be
valuable for theoretical model development and the calibration of a
model for specific operational use.  For example, in the simple model
now used for drift plumes, the effect of drift water solute concentra-
tion on the rate of evaporation from droplets has been neglected.
Incorporation of this effect requires a numerical solution of the drop
evaporation equation and increased  charges for computer use, which may
or may not be justified.  With model-to-graphics interaction, only a  few
iterations would be required to display deposition flux rate vs. distance,
with and without evaporation assumptions,  in enough cases to assess the
significance of the proposed improvement to the model.

-------
5553
                                                5     *908   saee    eeee

                                                   DISTANCE FROH  TOMER GO
    ADDITION TO RELATIVE HUMIDITY
             AT GROUND LEVEL

   .993
(9
o

I
a
             i   I  r
                                                                                                                                  I
                                                                                                                                  u;
                                                                                                                                  00
                                                                                                                                  I
         DISTANCE FROM TOWER Cm)
                                                    e    Tee     T5e     255

                                                     DISTANCE FROM TOWER CrO
                                                                                                                      255
                                                                                                                            .3««
             Figure  17.   Simulated behavior  of a vapor plume for  a natural-draft cooling tower,  typical

-------
                              -39-
                                                   SCALE

                                                    200m
Figure 18.  Sample contours of yearly drift  deposition  in the vicinity
            of two natural-draft cooling towers.

-------

                                                                                                             o
                                                                                                              i
Figure 19.  Three-dimensional representation  of yearly drift deposition in the vicinity of
            two natural — draft  oooliriR  towor-s.

-------
 0
 L
 .0
 L
 0
 UJ
 h-
O
H
I-
H
W
O
Q.
LI
Ci
     3088
     2888
     1888
        8
          8
                      580          1088         1580         2888


                              DISTANCE FROM  COOLING  TOWER  Cm)
                                                                        2580
                                                                                     3888
Figure  20.  Plot of yearly drift deposition vs. distance from two natural-draft cooling

           towers.

-------
                               -42-
Air Flow Behavior Patterns

     Other types of model predictions can be depicted by computer graphics.
Figure 21 shows the predicted pattern of air flow around a building by
using a computer model and graphics developed by Dr.  T.  J. Crawford,  Air
Quality Branch, Division of Environmental Planning.   Flow is  shown in
the vertical plane.  The numerical model that was used is based on the
Navier-Stokes equations, which balance the shear pressure forces with
momentum in combination with conservation of mass.  Each arrow is a
vector, with its length proportional to air velocity and its  direction
the same as the direction of air flow at that particular point.  Note
the "cavity" recirculation region that forms on the  downwind  side.  This
region may be enlarged on the CRT as shown in Figure 22.  Because the
building is about 100 feet high, a point of gaseous  emission  must be at
an elevation of at least 150 feet to preclude entrainment in  this "cavity."
Figure 23 shows the simulation of flow patterns around two mechanical-draft
cooling towers; various planes and cross sections are shown.   Obviously,
numerical tabulations of the same results would be much more  difficult
to interpret.

Future Demonstrations

     There is great potential for numerous future applications of computer
graphics to analyze air quality.  Some of the demonstrations  planned
will use other types of computer graphics hardware.   Refresh  graphics
will be used to dynamically manipulate a display on  the CRT.   For example,
a series of computer simulations of plume behavior with hourly variations
in meteorological conditions can be conveniently condensed into a brief,
meaningful, animated sequence.  This type of use is  expected  to increase.
Software development will stress the practicality of interactive analysis
to permit integrated assessment (e.g., consideration of the costs related
to pollution control alternatives and either simple  or complex environ-
mental consequences).  One system that has been considered would use
existing capabilities of TVA, EPA, and others to model fixed  and variable
costs of pollution control processes, residual generation, transport of
atmospheric pollutants, and ecosystem impacts.  Once collected, these
models could be incorporated into a flexible data base and analysis
system that would permit rapid exploration of alternatives for siting,
process design, and state of exogenous system (i.e., meteorology to
evaluate environmental consequences).  Where the state-of-the-art of
modeling is not fully developed (e.g., impact damage functions),
empirical or assumed relations could be used.  The analysis scheme would
be sufficiently general as to be able to incorporate future models
readily.  (Although institutional problems will probably delay develop-
ment of broad-spectrum, integrated environmental assessment indefinitely,
reductions in computer costs do make such ambitious models more feasible
now than ever believed possible.)

     A typical analysis session would be conducted interactively with
the computer.  During 'each step, various analysis options would be
presented to the engineer.  These options might include the ability to
(1) select alternative analysis procedures and graphic displays,  (2) apply
increasingly complex models, or (3) input different exogenous  conditions
or design variables.  Output of the results would be displayed in  tabular.

-------
SCALES,   SPACE-  185.7833M/IM   VELOCITY"  18.OTCM/S/HO
                                                                                                                 U>
   Figure 21.  Simulated pattern of simulated air  flow in the vicinity of a building.

-------
                                                                                                            I
                                                                                                            .p-
                                                                                                            -p-
                                                                                                            I
Figure 22.  Enlarged view of  simulated  entrainment cavity on downwind side of building.

-------
                                 -45-
      SYMMETRY PLANE
                         \\\\\ \N\ \\\\
                                       '\ \ N. N V
— •
— *
— ^
» — *
*
• •*
„

^v!
«^\
r)
\
y
)
\ * «
\ t *

* +
/ f
/ X
/ s
:
,
X
^
— *-
_*
«

•

•-- X
V \

^. * *
\ 1 '
\ * '
! : ;
-
,
s
       OROUND
       PLANE
                             X ^ -^
                                     "^X X
              \
              u--J
                                 /

                                 s
                                                              UP
      OLOSS SECTION *-A
      *    /   *
»\
- \}
                                             CROSS SECTION
Figure  23.  Various views of  simulated air  flow in the vicinity of
             two mechanical-draft cooling  towers.

-------
                               -46-


map, or graphical  format on a CRT or, at the user's option, directed to
a  computer plotter.  Statistical operations (e.g., ranges, confidence
intervals, regression analysis) may be advantageous for some analyses.

     The data base structure for such a system could include several
features:

1.   A data base containing a set of default variables necessary to run
     the various models for residual generation, environmental transport,
     and impact.  Any of these variables could be changed interactively
     through input programs.

2.   A regular, gridded, spatial data base for source location, land
     use, and topographic information.

3.   A costing data base relating to economics and performance models
     for control technology.

WATER QUALITY ANALYSIS

Data Display

     Figures  24 and 25 show typical plots of aquatic data used in reports
for TVA's Thermal Effects Studies for the Federal Water Pollution Control
Act, Amendments of 1972, Section 316(a).   The total time to produce
figures of this type alone over the life of TVA's "316" studies was
reduced from an estimated 176 man-days,  if manual techniques had continued,
to 8 man-days with the use of interactive computer graphics.

     Three-dimensional (one variable as  a function of two others) presen-
tations of water quality data can be valuable.  Demonstrations developed
to date can be grouped into three categories:   (1) horizontal spatial
distribution of a particular biological,  chemical, or physical variable
(Figure 26),  (2) variation of a water quality with distance and depth
(Figure 27),  and (3) variation of a variable with location and time
(Figure 28).

     In the future, other types of three-dimensional representations of
complex relationships will be used increasingly.  For example, De Angelis
and Thorp15 recently published computer program documentation for visua-
lizing multiple surfaces in space.  They report applications to ecosystem
studies such as three-level food chain systems, niche theory, and optimi-
zation strategies.

Interactive Analysis of a Water Quality Data Base

     A combined package of graphics and analysis routines was developed
in association with this project by TVA's Division of Environmental
Planning, Water Quality and Ecology Branch, to evaluate trends in the
quality of discharged water for TVA's reservoir system.  Weekly average
values for the volume of water discharged, reservoir depth, concentration
of dissolved oxygen, and temperature of the discharge are stored on the
computer for each of TVA's reservoirs during the last 16 years.  The
user can request a display of the weekly data for the entire period of

-------
o
H
I—
<
 TOTAL  NUMBER AND BZOMASS


   scjrt CBIOMASS IN  mO/rrf5


      2468
	I	i	|	|
                            i
                            ae
       e         10         ae         30          40

         *e,ri  CNUMBER  PER CUBIC METER3
                              L
          ROTIFERA    CLADOCCRA   COPCPODA    HOfMSS
                                                                              TEMPERATURE
                                                                     49-.
                                                                    aeee
                                                                           FLOW  INFORMATION
                                                                      i

                                                                      5678

                                                                    DAY OF MONTH CDEC)
       Figure  24.  Typical display of biological and physical water quality data  for TVA-EPA

                  Thermal Effects Studies.

-------
 SURFACE
                                                            1 - m  DEPTH
   345
    STATION
3 - n  DEPTH
   345
    STATION
                                PHYTOPLANKTON
                                CHLOROPHYLL a
                  xie"
5 - »  DEPTH
  345
   STATION
                                                                            X18'
                                                                                        oo
                                                                                        I

-------
 o
 IT

 Z
 LJ
 O
 z
 o
 o
 CD
 >-
 X
 o

 o
 UJ
 en
 Q
1067
IOS
104
103
102
101
100
99
                                    HOLSTON RIVER MILE
     Figure  26.  Three-dimensional  representation of distribution of dissolved oxygen in the

                vicinity of a dam  and a thermal discharge.

-------
              NOVEMBER  10, 1976
                                                    444
                                  443

               44Z             occ  O,K/PR  MILE
                        -TENNESSEE
                                                                      445
                                                                                       I
                                                                                       Ul
                                                                                       o
                                                                                       i
Figure 27.  Three-dimensional  representation of turbidity as a function of river mile
          and deptH.

-------
                                                     I INI
                                                                        Q LJ
                                                                                   • Y
I
     90 r
i
Ul
     Figure 28.  Variation in the National Science Foundation's water quality index as a

                 function of time and Tennessee River reservoir.

-------
                               -52-
 record  or  for  an  individual year.  Figure 29, for example, shows a
 typical plot of weekly average concentration of dissolved oxygen dis-
 charged from Fort Loudoun Reservoir.  The user can also request that
 yearly  average values for a particular variable be calculated and plotted.
 Figure  30  shows the yearly average values for dissolved oxygen discharged
 from Fort  Loudoun Reservoir.  The user can request computation and
 display of the cumulative amount of dissolved oxygen required to raise
 the average discharge value to a specified level.  These results show
 the change in  biological degradation (oxygen demand) in the reservoir.
 Figure  31  shows such a plot for Fort Loudoun Reservoir.  Note that, in
 1967, there appears to have been a marked reduction in the amount of
 dissolved  oxygen  needed to reaerate the discharge.  Finally, information
 from various reservoirs can be combined to display trend plots, as shown
 in Figure  32, which illustrates spring temperatures in the major Tennessee
 River reservoirs  over the past 16 years.

Water Source Mapping

     A system was designed to assist in the preparation of graphics and
 reports required  for reports on siting nuclear power plants and environ-
mental impact statements.   A map is needed to show ground and surface
water supplies, both municipal and industrial, within a 20-mile radius
of proposed plant sites.   A listing of the use characteristics for each
source must also be prepared.   This information is updated periodically,
and a new map is prepared about eight times during the assessment process
 for a typical nuclear power plant.   Manual techniques for recording the
 information and preparing the displays are time-consuming and tedious.

     A system was designed around a data  base containing information on
water use  and source locations associated with proposed sites.  Programs
were identified to generate specially formatted lists and special graphic
displays.   Updating routines were provided to keep the data base current.
The procedure  for generating a typical display involves several steps:

 1.   All the data or a particular subset of the information on water
     source can be called from the data base for plotting.

 2.   A  symbol  characteristic of each source type with an identifying
     number is plotted on the terminal screen at a position corresponding
     to its exact latitude and longitude.  The plant site and a
     representative 20-mile circle are also drawn (Figure 33).

 3.   Source symbols and identifying numbers will commonly overlap,
     resulting in a confused and unclear display.  Using the terminal's
     cross-hair cursor, a particular source symbol can be identified
     to the computer and adjusted to a new location.  The process
     continues until a satisfying arrangement results (Figure 34).

 4.   The final desired positions of all symbols are then stored.  A
     plot  is generated on a flat bed plotter using the stored symbol
     locations with the size adjusted to a particular scale (e.g.,
     1:250,000).  The results then are transformed photographically
     into  a clear acetate overlay, which can be placed over a standard
     base  map  and photographed to arrive at the final figure for the
     report (Figure 35).

-------
                                                                                                                Ul
                                                                                                                I
      1969 1961  1962  1963  196-4  1965  1966  1967  1968  1969 197« 1971 1973 1973  1974  1975  1976  1377
                                                YEAR
Figure 29.   Average weekly concentration of dissolved  oxygen in discharge waters  from
             Fort Loudoun Reservoir over the past  16 years.

-------
  ^  6
CD


0
          1 I I I I I I I I I I . I I I I  i I I  i i i I i i i i I i.illiilililillii'tli.ii I i . , I ! . I i I 1 i I I  l 1 i I ; 1 I .
      1960 1961  1968  1963  1964  196S 1966  1967  1968  1969  1970 1971  1973  1973  1974  1975  1976

                                                YEAR
                                                                                                               l
                                                                                                               Ol
Figure 30.  Yearly average concentration of dissolved  oxygen in discharge waters  from

            Fort Loudoun Reservoir over the past 16 years.

-------
   69999
—  49999
   29999  -
                                                                                                                I
                                                                                                                Ul
                                                                                                                Ul
                                                                                                                I
        1968
                  1965
196-4
1966
                                                 1968
                                                 YEAR
197*
                                                                      1972
                                                                                1974
                                                                                           1976
Figure 31.  Yearly average amount of oxygen  that would have to be  added to discharge
             waters of  Fort Loudoun Reservoir for the past 16 years  to  bring the
             concentration to the specified levels.

-------
KENTUCKY
   PICKWICK
        WILSON
          WHEELER
        GUNTERSVILLE
 4*.           NICKAJACK
  * >           CHICKAMAU6A
                      WATTS BAR
        y-             FT. LOUDOUN
        Sf*
                                                                YEAR
                    OUTSIDE BOUNDARY INDICATES  10" C
Figure 32.   Average  temperatures of spring discharge water for  reservoirs on the
            Tennessee River  over the past 16 years.

-------
                              -57-
Figure 33.  Initial display by cathode ray tube of water source
            symbols plotted on the basis of latitude and longitude.

-------
                               -58-
Figure 34.  Final display by cathode ray tube of water source symbols
            after visual adjustment.

-------
                            -59-
                                                  •--—- l-Jr_^_-- -™—


                                                   "°/Ah«.,  A  /  | -*
                            • ~®  - 1 '   O Houston
                            ' N/' HMNAKFORKSr, f
                  	  ' "•  ' 'N   '')'/

               COURTLAND NUCLEAR PLANT SITE

                              0 SITE


          WATER SUPPLIES WITHIN A 20 MILE RADIUS OF PLANT SITE

            D  PUBLIC SURFACE       & PUBLIC GROUND


            A  INDUSTRIAL SURFACE   O INDUSTRIAL GROUND
Figure  35.  Figure for  final report prepared by interactive graphics
           showing water sources in the vicinity of  a proposed nuclear
           electric generating facility.

-------
                               -60-
     Although  the advantages of using a computerized graphics display
 system were demonstrated, the methodology was never implemented on a
 routine basis  because of the limited accessibility and slow response of
 TVA's overloaded computer system.

 Interactive Analysis of Water Quality Models

     Numerous  mathematical models of water quality are used routinely
 for environmental analysis.   These range in complexity from modifica-
 tions to the classic Streeter-Phelps equation for a deficit of dissolved
 oxygen in streams to those models that predict detailed spatial distr
 ibutions of multiple biological, chemical, and physical variables in
 lakes, reservoirs,  and estuaries.  Many of these models can be used
 interactively  in a time-sharing computer environment.  Others require a
 great deal of  computation necessitating batch processing.  Both input
 entered into and output generated by these models can be complex and
voluminous.   For example, one of TVA's water quality models can predict
 the concentration of 18 variables in a deep reservoir at each meter of
 depth, four times a day, during the course of a year.16  Input to this
model consists of hydrodynamic, hydrologic, chemical, biological, physical,
 and meteorological conditions that impact on the reservoir.

     Another model commonly  used predicts temperature distributions in a
 reservoir.   Input data consist of daily inflows and discharges, solar
 radiation flux, cloud cover, wind velocity, and temperature of the
 inflow water.17  Also, a number of variables are required to define
 hydraulic mixing, water clarity, and reservoir morphology.   This model
 requires about 30 seconds of CPU time on TVA's computer system to complete
 a one-year simulation.  An interactive scheme of input and analysis was
 developed (Figure 36).  Overall program control and file allocation are
 embodied in a  command language program (CLIST).  Each of the various
 steps of the analysis can be accomplished by answering yes-and-no ques-
 tions.  The user must, however, be familiar with the features of this
 model since the intent of this demonstration is to provide an analysis
 tool to extend the technical capabilities of a water quality modeler,
 rather than provide modeling capability to someone unfamiliar with this
 particular analysis.

     A typical terminal session involves several steps:  The engineer
 logs on the time-sharing system and selects a certain data set suited
 for the model.  This data set might have been generated initially with a
 preprocessor program designed to interface with various types of data
 input sources  and transform that data into a format required by this
 particular model.  If this simulation is to be compared with a previous
 result, the file in which this information is stored is identified.
 Next, the user has the option of interactively changing various input
 variables in the model.  When this is completed, the model can be run
 for the desired simulation period.  All output is directed to a storage
 file with direct access for analysis and display.  Statistics concerning
 the accuracy of the model in terms of actual field data are calculated
 for the entire water column.  Other statistical comparisons between
 predictions and field data could also be made.  For example, the user
 may need to know how well the model predicts average temperatures of

-------
                               -61-


epilimnion, hypolimnion, surface, or discharge waters.   For this demonstra-
tion, however, overall simulation accuracy of the predicted temperature
profile was selected for illustration.

     After the simulation run is successfully completed, the user is
presented with a menu of possible analysis options (Figure 36).   Profiles
of temperature vs. depth are of primary interest.  The user may select
the display of any profile based on its Julian day number, and profiles
may be overlaid on each other.  Figure 37 shows a typical display.
Results from a previous run may also be displayed.  When the user has
satisfied himself with this portion of the analysis, he may return to
the original menu of display and analysis.  The user may then decide to
display the predicted temperature profile overlaid with the profile
actually measured for that day in the field.  Figure 38 shows such a
plot.  The user may then select for display the statistics for the
simulation.  He may choose a tabular listing of the days on which field
data were collected and the resulting simulation statistics.  These may
also be displayed in graphical form (Figure 39).  The left graph in
Figure 39 shows the root mean square (RMS) for the simulation; the
magnitude of this value is the average, unsigned deviation of the pre-
dicted results as compared with actual field data averaged over the
entire water column:
     RMS =
             IN
             I  (Calculated - Observed)
                          N
where N = number of observed points.  For this simulation run, the plot
shows that the predictions are less accurate in the spring and fall than
in the summer.  The right graph in Figure 39 is a plot of the average
mean error.16  This statistic is the product of the average temperature
of the water column and the normalized mean error (NME), which is
calculated by:
               ,- (Calculated - Observed)
               i.         _.,      •        X
                         Observed
     NME =   —	   	
                             N

     It provides to the engineer an indication of the relative magnitude
of the deviations of predictions from observed results.  Thus, Figure 39
shows that the model predicts lower values for average temperature in
the spring and fall and higher values in the summer.  When the analysis
is completed, the user returns again to the basic menu of display and
analysis.

-------
                               -62-
     One final type of analysis could commonly be made.  For this, a
display could be generated of selected input data such as meteorological
or hydrological conditions or secondary model output data such as the
average temperature of water discharged from a reservoir.  This data may
be viewed over the entire simulation period or over a selected time.

     When these analyses are completed, the user may return to the
interactive input routine, making necessary changes in a variable, or
terminate the session.  Simulation results may be stored in a data set
at this time for later retrieval.   The use of this analysis capability
has greatly enhanced the ability of water quality modelers to apply this
particular model to specific situations.   Previously, processing of
model output was accomplished manually with, in some cases, days elapsing
between simulation runs.   A sensitivity analysis to determine the model's
response to small perturbations in input variables was exceedingly
time-consuming.   The work can now be accomplished conveniently in much
shorter time.

Future Applications

     Computer-generated animation of highly dynamic phenomena will be
used more widely,  particularly in studies of hydraulic mixing, biological
population migration (e.g., fishery studies), and biological succession.
Computer graphics will provide the means  for interacting with systems of
environmental analysis models and broad data bases.  For example, to
quantitatively assess the impact of strip mining on aquatic ecosystems,
models can be used to evaluate the (1) physical and geological charac-
teristics of the mining site, (2)  mining techniques used, (3) reclamation
techniques employed and results expected, (4) hydrology of the area, and
(5) ecosystem impacts.  These models will eventually be capable of
integrating the effects of multiple mining sites at any point downstream
from these operations.  After changes in the quality and quantity of
water are translated into biological and aesthetic impacts, they can be
assigned costs or semiquantitative rankings.  Practical application of
such an analysis capability demands that a means be provided to facilitate
data input, component model interaction,  and output display.

RADIOLOGICAL HYGIENE

     The analysis of existing and potential impacts from radiological
material can be grouped into two categories:  (1) engineering studies,
which include modeling the environmental transport of nuclides, protec-
tive shielding studies, and expected dose calculations; and (2) analysis
of environmental monitoring and personnel exposure data.  Applications
to date have concentrated on the visual presentation of data.  Samples
of these displays are described below:

1.   Dose conversion factor (to convert a measured air radioactivity to
     dose rate when given a particle size distribution) vs. particle
     size for uranium-238 inhalation dose to lung (Figure 40).

2.  ' The probability of a given radiation exposure based on field measure-
     ments; a linear plot is characteristic of background (Figure 41).

-------
                                                        RESERVOIR TEMPERATURE

                                                         PROFILE  SIMULATION

                                                              ANALYSIS
                                                           MODIFY  DEFAULT
                                                           INPUT DATA SET
                                                           RUN  TEI1PERATURE
                                                           SIMULATION MODEL
                                              DISPLAY A'lD ANALYSIS OF SIMULATION  RESULTS
                                                                                                               OJ
                                                                                                                I
 DISPLAY TEMPERATURE
      PROFILES
      COMPARE
    V/IT1! FIELD
       DATA
    STATISTICAL
    ANALYSIS OF
    SIMULATIO.'!
   DISPLAY  I.NPUT f.
  SLC"NDA"Y Ol'TPI'T
        DATA
      CO'ITIN'UE
Plot profile
Overlay  profile
erase screen - new profile
Overlay  profi le of
previous run
Overlay  predicted
and actual data
Erase  screen - select
new day
Listing of statistics
Graphics of statistics
Discharges
Inflow volunfi
Inflow temperature
'find velocity
Isolation
Reservoir surface
ol
-------
                             -64-
              50'
              40-
           I  30-
              20-
              10-
                          TEMPERATURE CC>
                            10         28
                    till
J	I
                   30
J	I
                             1    /
                        '    /    /    KEY
                           1    / DAY 89   	
                           '    /  DAY 175  	
                           r    ,'   DAY 128	
                               |   DAY 258  	
Figure 37.  Profiles of predicted temperatures and depth selected
           interactively by the user and displayed on a cathode ray
           tube.

-------
                               -65-
                    0
TEMPERATURE  CO
  18         20
                50-
                40-
            1  30


            UJ
            Ul
                20-
                10-

                                     I  i
                                              i  i  i
                                   SIMULATION FOR PAY 114
                                             OBSERVED
                                            PREDICTED
Figure 38.  Profile  of predicted temperature and depth overlaid with
            actual temperature profile measured in the field.

-------
    1.00
 LJ
LJ
    0.50
o
°  0 K
ty  *'25
    0.00

                    ioe        we
                        JULIAN DAY
400
                                                              -6
                          100        see        300
                               JULIAN  DAY
  !
400
         Figure 39-  Plot of  statistics (root mean square and average mean error) calculated over
                     entire water column for a simulation run.

-------
      10
        16
QC
o
10
   15.

-------
  §
CO O
O K

X O
W 0*1
  0
  H
  S
                                                                                                             00
                                                                                                              I
                                      Z- VALUE (PROBABILITY)
Figure 41.  Probability of a given radiation  exposure based on field measurements.

-------
                               -69-
3.   Curves of time vs. distance for a given exposure to the thyroid
     gland from an accident situation (Figure 42).

4.   Gamma-ray spectrum of cesium-144, as measured by Ge(Li) detector
     (Figure 43).

     Future applications include interfacing interactive graphics with
mathematical models of the environmental transport of radionuclides.
Computed airborne dispersion patterns can be combined with population-
activity information to produce overlays on a CRT to evaluate deposition
and dose effects under a variety of conditions.  Graphics combined with
exploratory statistical procedures can be used to identify trends in
environmental monitoring data.  Graphics will inevitably play an integral
part in the implementation of a nearly real-time analysis system for
predicting the transport of accidental releases of radioisotopes to the
atmosphere or receiving waters.  After the necessary input conditions
are specified, automated procedures will compute dispersion patterns.
Graphics will be used to interpret and communicate these results rapidly
so that suitable monitoring and hazard abatement schemes can be implemented.

INDUSTRIAL HYGIENE—DISPLAY OF NOISE DATA

     Measurement and control of noise in the working environment presents
opportunities to effectively use computer graphics.  Octave band field
measurements of noise levels can be displayed as shown in Figure 44.
Curves of standard frequency attenuation for various types of personal
protective devices can be applied to this curve to show the noise levels
expected to reach the ear for different frequency bands.

     Plans of TVA's Industrial Hygiene Branch call for developing a
capability for modeling noise behavior in the immediate working environ-
ment and in the community.  The ability to change source locations
rapidly, to visually position and alter the types of attenuation barriers,
and to investigate potential impacts on land use and population distribu-
tion can be accomplished most effectively through interactive computer
graphics.

SOCIOECONOMIC IMPACT ANALYSIS

     Development of a methodology that uses computer graphics for screen-
ing potential sites for power plants according to socioeconomic criteria
has been underway for two years.  During the first year, a review was
conducted (by"a consultant) of the state-of-the-art of identifying and
measuring socioeconomic impacts of large-scale construction of power
plants.  TVA's procedures for analyzing socioeconomic impact and miti-
gation were also reviewed.  Those capabilities of a computer graphics
analysis system that were needed to assist with these analyses were
identified.

     Twenty-two counties in East Tennessee were selected for testing a
screening methodology.  Data were collected on 24 socioeconomic indi-
cators (Table 4) and incorporated in a data base that could be manipu-
lated by interactive routines.  Basic data management, analysis, and
display routines are listed in Table 5.

-------
            183-
            182-
        "   18-
        H
           18
            rl.
               102
                           26 rw
I    I  I  I  I I I I
                        I  I  I  1 I I I
             103                 18'

                  DISTANCE CiO
                                            1  I  1  I I I I
185
                                                                                                         o
Figure 42.  Plot of time vs.  distance for a  given dose  to  thyroid gland for an accidental
            exposure situation.

-------

       d
       Of
             ZS9
                             260
                                             27»              280



                                                ENERGY  Ck*v>
                                                                            29*
                                                                                            3*0
Figure 43.   Gamma-ray energy spectrum for  cesium-144 as measured by a Ge(Li)  detector.

-------
     185
     109
                                                O-UITHOUT EAR PROTECTION        A-UT- 1*9.48
                                                A-EXPECTED LEVEL UITH EAR PLUGS A-UT-  83 82
                                                a-EXPECTED LEVEL UITH EAR HUFFS A-UT-  75.31
                                                                                                                   ho
                                                                                                                   I
                                                                        19
10
                                                   FREQUENCY
Figure  44.   Display of octave  band noise measurements  and predicted noise levels when
             •personal, protective equipment  -is usecl .

-------
                         -73-
TABLE 4.  COUNTY-LEVEL SITE SCREENING INDICATORS
          TESTED FOR IMPACT ANALYSIS
 1   Recreation:  recreation acres per capita
 2   Health:  population per physician
 3   Police:  expenditures per capita (weighted average of
              each county with its cities)
 4   Expenditure per pupil
 5   Percent students overcrowded
 6   Average teacher salary
 7   Number of pupils per teacher
 8   Percent of county population in jurisdiction having a
       planning commission
 9   Percent of county population in jurisdiction having a
       comprehensive plan
10   Percent of county population in jurisdiction having a
       zoning ordinance
11   Percent of county population in jurisdiction having
       subdivision regulations
12   Percent of county population in jurisdiction having
       capital budgeting
13   Percent of county population in urban places
14   Housing vacancy rate (1970)
15   Percent of housing built before 1950
16   Percent of change in population (1960-1970)
17   Percent of change in population projected  (1970-1990)
18   Unemployment rate
19   Median family income (1970)
20   Percent of population receiving welfare
21   Percent of population in 20- to 44-year age group
22   Percent of population living in same house five years ago
23   Percent of population served by public water
24   Percent of population served by public sewer

-------
                                -74-
TABLE 5.  CATALOG OF DATA BASE MANAGEMENT, ANALYSIS, AND GRAPHICS ROUTINES
          FOR A COUNTY-WIDE SOCIOECONOMIC INFORMATION SYSTEM
  Name
                      Function
CALLO


CFILL

CEDIT
   Data base management

File allocation routine (to allocate data
  files to proper logical units)

To fill data base with multiple data values

To display and alter individual data values
CFIND
CRANK
     Analysis routines

To search the data base or a subset of
  counties for a given data value condition
  (less than, greater than, equal to)
To
» prepare basic statistics and ranking of
counties for a given variable
CSUM
CXYPLT
CKIV2
CDRW
          Display

To prepare a formatted list of data values
  of an individual county or a group of
  counties

To prepare a scatter plot of two county
  variables (one variable against the other)

To plot up to 12 Kiviat diagrams with a
  maximum of 20 axes

To plot selected county boundaries and one
  data value at the center of each county

-------
                               -75-
     Selected indicators were combined mathematically to form capacity
indexes, thus reducing the number of indicators to manageable size and
providing a better measure of the potential for each county to absorb or
benefit from a particular type of impact.  Graphical and statistical
methods were used to evaluate the relative merits of each new index
formed.  Resulting from this work were six indices:  (1) public service,
(2) planning and public administration, (3) health, (4) education,
(5) growth absorption potential, and (6) economic need.  Finally,  several
forms of a composite index composed of these six capacity indices  were
tested.  Procedures were also developed to weight various indicators and
capacity indices in these analyses.

     Three types of graphic displays were used in this research.   Figure
45 shows the results of a program that draws any or all of the county
boundaries being considered, identifies the county, and places a value
or representative symbol within the boundary.  The map can be displayed
at various scales, permitting a small selected portion to be enlarged.
Figure 46 shows six socioeconomic indicators displayed in the general
form of the Kiviat diagram previously described.

     Figure 47 shows another useful graphics display.  Scatter plot dis-
plays were generated for pairs of indicators whose correlation coefficients
were unexpected or otherwise of interest.  The entire set or a given
subset of county data for the selected indicators  can be plotted readily,
and the axes labeled.  Details of this work are presented in Appendic C.

GEOGRAPHIC INFORMATION SYSTEMS AS AIDS TO SITING FACILITIES

     Two research tasks were undertaken:  (1) to review the state-of-
the-art with respect to the use of computer-assisted geographic infor-
mation systems to support the process of siting major power generating
facilities and transmission corridor routes and (2) to provide suggested
design criteria for planning a geographic information system for siting
power plants.  The results are summarized below; details are available
in separate reports.19'20

     To accomplish the first task, a survey of siting methodologies was
conducted among public and private utilities in the United States and
Canada to determine whether geographic information systems are used and,
if so, for what specific siting phases and analysis processes.  Also
included in this survey were those State governments that maintain
geographic data systems that might be applicable to siting power
facilities.

     The methodologies most frequently used by utilities and their
consultants to site power facilities include checklist, overlay, and
matrix techniques.  In general, utilities have largely restricted their
siting efforts to developing comprehensive checklists of siting consi-
derations, whereas consulting firms have concentrated on developing
techniques (referred to as methodologies) for evaluating a subset of
considerations (such as environmental impacts) at  a particular stage of
site selection.

-------
                                                                                                           I
                                                                                                          —I
Figure 45.  Display of county boundaries in the socioeconomic methodology test area.

-------
                PERCEMT POPULATION
                      URBAN
            HOUSING VACANCY RATE
                 AVERAGE TEACHER'S
                      SALARY
                                                     PERCENT LIVING IN SAME HOUSE
                                                            LAST 5 YEARS
                                                   --  PERCENT HOUSING PRE-1950
                                                     POPULATION PER DOCTOR
COUNTY A
COUNTY B
COUNTY C
COUNTY 0
COUNTY E
COUNTY F
      Figure 46.  Sample of  socioeconomic indicators displayed as Kiviat  diagrams.

-------
                            -78^
1UW — I 	
J
i
1
1 A
z
z
H//V (^ rt
WJ oW
> UJ
HO
ZtL
O
HZ 40
1— <
_JQ:
DD
(L
O
ft
-






A
^ A
A
A :
A

A A A A
A
A
A
1
111^ -"p 1 **| I i i ; | T1 1 i [ I
7000
8000         9000         10000

     AVERAGE  TEACHER'S
           SALARYC$>
11000
Figure  47.  Scatter plot of percent of population living in urban
           setting vs. average teacher's  salary.

-------
                               -79-


     The most popular approach at all stages in the siting process is
the checklist technique.  Overlays are being used increasingly at stages II
(multicounty) and III (site-specific) to graphically depict specific
sets of variables such as engineering constraints or environmental
impacts.

     There are four basic methods for routing transmission lines, ranging
in sophistication from a direct line (i.e., a line is projected from
source to load) to computer-assisted optimization models (i.e., the
probable costs and benefits of a wide variety of alternatives are analyzed,
and the routes are selected and ranked according to impact).  At present,
most utilities rely heavily on direct-line routing.

     Computer-assisted geographic information systems are not commonly
developed or used for siting or for routing transmission corridors.
Efforts on the part of utilities and consulting firms to use geographic
information systems for routing transmission lines have encountered
various difficulties:

1.   The level of detail and the degree of resolution of a system's data
     base usually is inversely related to the amount of area under
     consideration; detailed, finely resolved, expensive data bases are
     required for siting transmission lines.

2.   The cost ratio between siting and constructing lines becomes unfavor-
     able as data capture expenses for the geographic information system
     increase the cost of siting.

3.   Tradeoffs between resolution and data capture expenses have been
     made without the benefit of prior experience.  In many instances,
     the resultant analyses have not been useful because coarse resolu-
     tion eliminated potentially acceptable areas.

     One of the most prevalent difficulties associated with assessing
sites for power generating facilities is that of analyzing and evaluating
large amounts of incommensurate data collected and organized by many
diverse sources.  Although computer systems have the potential for
becoming a useful tool for analyzing these siting variables, they have
not yet been developed for or used by utilities.  Geographic information
systems exist in various forms and degrees of completion.  They differ
in form of data storage structure, and degree of computer assistance
that can be provided to the user, but in many respects they are quite
similar.  All the systems use computers to store data, and most can
retrieve information in map format.  Most are structured for some type
of natural resource planning and include topographic data, soil type
distributions, and land use information.  Also, data concerning aquifer
recharge areas, unique and endangered species, natural features, historic
areas, unique and endangered species, natural features, historic areas,
population distribution, and climate are frequently included.  Most of
the systems use a regular grid format for data analysis.  Data are most
frequently encoded directly to cellular format, but the trend is toward
digitizing information in terms of irregular polygons and using the
computer to convert to appropriate cell sizes for analysis.  Resolution
(cell size) varies widely among systems.

-------
                               -80-
     Statewide computer-assisted geographic information systems devel-
oped for planning purposes are generally not suitable for direct use by
utilities for siting.  However, they appear to be useful for providing
information pertinent to licensing power plants.  Geographic information
sharing between utilities using these systems could lower the time and
cost of data collection.

     After the use of geographic information systems for siting power
facilities was surveyed, a study (Task 2) was undertaken to determine
general design considerations of a computer-assisted geographic infor-
mation system that potentially could aid the assessment function within
a generalized, comprehensive methodology for siting power plants.

     A hierarchical set of subsystems, processes, and modules that
compose such an information system was identified.  Estimates of
resource requirements for system development and implementation were
summarized for each subsystem and data base.

     A geographic information system can be subdivided into subsystems,
and individual subsystem components are properly defined by a thorough
investigation of user information needs and objectives.   Five generic
subsystems can be defined:   (1) the system management subsystem, consist-
ing of the system maintenance, scheduling, and operating procedures, the
user-system interface, and staffing requirements; (2) the data acquisition
subsystem, including procedures for compiling, cataloging, and filing
source documents, and techniques for digitizing (converting information
to a machine-readable form) data from maps, aerial photographs, or other
sources; (3) the data base management subsystem, which maintains and
provides procedures that control the access and use of the data base;
(4) the data analysis subsystem, including analysis capabilities such as
data transformation techniques, derived data processes such as index
construction, coincidence and proximity analysis, modeling, statistical
analysis, and others; and (5) the data display subsystem, which provides
procedures and devices to display output in the form of summary reports,
tables, or graphic displays.

     One additional aspect of a geographic information system, although
not strictly a subsystem, relates to the use of the information.  Proce-
dures need to be devised for effectively applying the generated informa-
tion to solve the problems of concern.  This will depend on the user's
particular management system for decision making.

     A computer-assisted geographic information system is a potentially
useful tool to help meet objectives of data handling, analysis, and
graphic display in a comprehensive methodology of power plant siting.
Possible design objectives of an information system and descriptions of
their relationship to power plant siting methodology are listed below.
Objective 1:  To store and retrieve spatial engineering, socioeconomic,
              and environmental data.

     The power plant siting process requires the use of large amounts of
geographically referenced information.  Certain issues related to computer

-------
                               -81-


storage and retrieval of information must be resolved:  the type of
location identifiers attached to the data, the coordinate system to be
used, and the hardware and procedures for encoding and storing data.

     A flexible system must be able to handle geometric representations
of data, including point, line, and area data.  Polygons could be used
as the primary data sructure for data storage, with the capability of
converting to regular grid cells for ease in manipulation.

     Location identifiers refer to referenced coordinates.   Latitude-
longitude can serve as the primary coordinate system, with the capacity
to transform into UTM, state plane, or other coordinate systems as
required.  Data encoding is the conversion of mapped information into
computer-readable form.  Particular encoding conventions depend on
geometric representations of the data and the desired digital structure.
Examples include predominant type, percentage, absence-presence, and
center point sampling.  Both manual and automated techniques for
accomplishing this task could be included in the system.

     An integral part of managing a large volume of data is the means of
data storage.  Provisions for both physical and digital storage techniques
are required in the system design, including manual files,  punch cards,
magnetic tape, and discs.  Each method has advantages and disadvantages
related to specific needs.

Objective 2:   To store data and analyses from past siting studies in
               a form that will allow easy retrieval and updating for
               use in future investigations in the same region.

     Data maintenance is an important consideration in geographic information
system design.  It includes data archiving, data base additions, and editing.

     The ability to archive historic data in a format that will facilitate
its use in future siting studies is a major factor in the cost-benefit
justification of the system.  A system should be able to accommodate
increased quantities of data and have the flexibility to add new spatial
data types.  Editing or error-checking mechanisms are needed to modify the
data base.  Data inaccuracies may occur as data are input or as a result
of an actual change in the condition as described at a particular location.

Objective 3:  To consider all locations within a region at a variety of
              levels in a scaled hierarchy from the multistate level down
              to specific sites.

     Most processes for siting power plants require collection and analysis
of data at several levels of resolution in a scaled hierarchy.  The criteria
for determining potentially suitable sites become more stringent as the
scale becomes more site-specific.  The system should be designed to accept
data collected at any level in the hierarchy and should be able to transform
data from one level to another.

-------
                               -82-
Objective 4:  To analyze the spatial distribution and interrelationships
              of environmental, socioeconomic, and engineering conditions.

Objective 5:  To review the effects of using a variety of siting scenarios
              that reflect differential weighting of siting criteria
              among engineering, socioeconomic, and environmental factors.

Objective 6:  To review potential areas for all types of generating
              facilities, including fossil, nuclear, and hydroelectric
              plants.

Objective 7:  To evaluate alternative development concepts such as
              nuclear parks.

Objective 8:  To consider future changes in siting and power generation
              technology.

     Objectives 4 through 8 address useful data manipulation and analysis
capabilities.  Possible system capabilities could include (1) rescaling
or restructuring of data values, (2) analysis of spatial relationships,
(3) logical combinations of data,  (4) mathematical manipulation, (5)
weighted indexing,  (6) statistical operations, (7) windowing, and (8)
special-purpose operations.

Objective 9:  To communicate  to managers,  regulatory agencies, and the
              public the information, analyses, and value judgments
              related to siting decisions.

     Effective communication  is an extremely important factor in the use
of geographic information.  Raw data and the results of analysis can be
communicated visually through computer graphics.   Information can be
presented as maps,  graphs,  or charts as well as tabular listings.  Computer
hardware output devices that  are most readily used for creating displays
include character printers, line and electrostatic plotters, CRT's, and
COM (computer on microfilm) devices.

     Two types of analyses, which fall under the general title of spatial
analysis, lend themselves to  interaction with a CRT.  Information used in
such studies is commonly referenced on a cellular basis.  Several well-
developed software packages such as SYMAP and IMGRID21 are available to
manipulate these data in a meaningful manner.  Two basic types of questions
can be asked:  (1) "Which cells have given attributes in common?", and
(2) "What is the relationship of a particular cell to its neighboring cells
or to other map features?"

     The attributes referred to in the first question may be the data
variables themselves or a combination of the raw data, which in turn
indicates a particular attribute.   Figure 48 illustrates a composite
plot generated by a line printer of septic tank suitability for Knox
County, Tennessee.   Factors considered in such an analysis might include
soil type, percolation rate,  and depth to water table.  In response to
the second question, a cell that is classified as having harvestable trees
might be a much better candidate for logging operations if it is surrounded
by other cells with harvestable trees, than would the same cell located in
a residential area or a park.  A similar question might be, "What cells
within a given distance from a major road or river meet certain criteria?"

-------
             KNOX  COUNTY
             TENNESSEE
Tennessee Valley Authority
Division of Forestry, Fisheries,
and Wildlife Development
     1975
       Source of data
^.^ J   County Soil Survey 1955
   i I   United States Department of Agriculture
    '   Soil Conservation Service
"•CALL
mile'j
                                                              SEPTIC  TANK
                                                              SUITABILITY
                                                               slight
                                                             I moderate
                                                               severe
                                                               water
                                                                                                                 CO
                                                                                                                 U)
                                                                                                                 I
             Figure 48.  Shaded  line-printer composite display of septic tank suitability for Knox County,
                        Tennessee.

-------
                               -84-
     To illustrate a typical interactive analysis that might be conducted
at a CRT, a test area composed of a 50 x 50 array of 2.5-acre cells is
selected.  Each cell was classified with respect to eight different
types of data (Table 6).  At the beginning of the analysis, the user
selects any three sets of data.  He would then request a display of any
one particular factor, such as developed land; Figure 49 shows a display
of those cells classified as developed land.  Other information related
to the area, such as political or property boundaries, roads and topog-
raphy, can be stored separately on flexible disc and then superimposed
on the screen; Figure 50 shows the roads and 50-foot contours overlaid
on the display.

     Next, the user might request a multifactor analysis.  In this
example, the user has requested a display of cells classified as being a
farm or an estate and having upland conifer trees and a land slope of
3 to 6 percent; Figure 51 shows the CRT display.  Cells having the three
conditions in common are shown with a star.  Cells that meet only part
of the requirements are shown with other symbols.  The program also
keeps count of the number of cells in a particular category.  By multi-
plying the count in each category by the size of each cell, one can
tabulate the number of acres for each group.  A bar chart of this
information is shown in Figure 52.

     This technique can be used, for example, to determine the number of
cells that remain available for a certain use.  Other similar analyses
that could be demonstrated relate directly to facilities siting:  location
of endangered species, proximity to water, avoidance of earthquake-prone
areas, or distribution of sensitive crops.  The analysis principle,
however, remains the same.

     The implementation of practical analysis techniques that use computer
graphics for regional environmental analysis (including facilities
siting) depends to a large extent on the development of a regional
geographic information system capable of accommodating a variety of
spatially referenced resource data.  As a result of the activities of
this research project and others within TVA, an interdivisional study
team made up of representatives from twelve TVA divisions was estab-
lished at the direction of the General Manager's Office.  Staff from
this research project participated heavily in this four-month (March-
June 1977) study effort.  Objectives were to identify (1) current and
future TVA needs for geographic information and (2) opportunities for
sharing geographic information and possibly developing the information
system.

     The first activity of this study involved compiling an inventory
for each TVA division, listing (l) program functions that use geographi-
cally referenced information, (2) a description of geographic information,
(3) data sources from which the information is derived, and (4) relevant,
existing, organized systems for handling this information.  A sample
inventory for the Division of Environmental Planning is shown in Table 7.

     A second task completed by the study team involved interviewing TVA
division management to identify current and projected needs for infor-
mation systems and geographic information.  A final task was the analysis

-------
       TABLE 6.  SPATIAL DATA  CLASSIFICATIONS  USED  TO  CHARACTERIZE  CELLS IN THE DEMONSTRATION AREA
Aspect—land orientation
  100% water
  Flat (less than 6% slope)
  North
  Northeast
  Northwest
  East
  West
  Southeast
  Southwest
Slope
  100% water
  1-3%
  3-6%
  6-10%
  10-15%
  15-25%
  25-45%
  Greater than 45%
Forest type
  Developed land
  Water
  Wetland
  Cropland or pasture
  Lowland coniferous
  Lowland deciduous
  Upland coniferous
  Upland deciduous
  Upland mixed
Depth to bedrock
  100%
  1-1.5 ft
  3-10 ft
  3-20 ft
  3-30 ft
  5-20 ft
  5-30 ft
  5+  ft
  100+ ft
Water—predominant type
  None
  Swales
  First-order streams
  Other streams
  Ponds
  Reservoirs
  Lakes
  Rivers
  Estuary
  Ocean
Residential land use
  None
  Farms and estates
  Single family—large lot
  Single family—medium lot
  Single family—small lot
  Multifamily--low rise
  Multifamily—medium rise
  Multifamily—high rise
i
00
 Commercial  and  industrial  land  use
   None
   Shopping  centers
   Downtown
   Strip-and roadside
   Wholesale storage
   Modern industrial parks
   Extractive industry
   Individual industry
   Old industrial complexes
Summary land use
  None
  Recreation
  Low-density residence
  Medium-density residence
  High-density residence
  Transportation
  Institutions
  Industry
  Commerce
Transportation—road type
  None
  Unimproved
  Paved, light-duty
  Paved, medium-duty
  Heavy duty
  Divided with access
  Divided with limited access
  Interchange

-------
                              -86-
         AAAA
         AAAA
         AAAA
                       A
                   A   A
                   AA AAA
                      AAAA
                      AAAA
                   A   AAA
          A  AA
            A AA
            AAA   A
                 AAA
                    A
                    A
        AAA
        AAA
           A
           AA
            A
         A A
          AA
           A

         AAA  AA
         AAA   A
          AAA

          AAA
          AA
          AAA
          AAA
          AAAA
          AAAA
          AAAA
             A    A
               AAA
                 AA
                  AA
                  AA
                   AA
                   AA
AA
AAA
   A
   AAA
    AA
                   AAAA
                   AAAAA
                   AAAAA
              AAAA  AAAAA
              AAAAA  AAAA
            A   AAA  AAAAA
            AAA   AA  AAAA
            AAA AAAAA AAAAA
            AAAAAAAAA AAAAA
        AAAAAAAAAAAAA AAAAAA
       AAAAAAAAA  AA A AAAAA
       AAAAAAAAA  AAAA  AAAAA
        AA       AA  AA  AAAA
                    AAAA  AAAA
                    A AAA AAAAA
                       AA AAAAA
                         A AAAA
                         A  AAA
                          A  AA
A                         A   A
                   AA     A
                 AAA
                AAAAA
                AAAAA
                 A  A
        A     A
        AA AA        A
       AAA           A A
     A              AAAAA
     A                AAA
  A  AA  AA           AA  AA
   AAAA AAAA           A   A
   AAAAAAAAA
  A AAAAA A
  A AAAAA
AAAAAAAAA                  A
 A  AAAA A           A   AAA
A   AAAA AAA     A   AAAAAA
A   AA  AAAA
        AAA    AA
A              A
A              A A
      AAAA
       AAA
        AA
         A
        AA
        AAAAA
        AAAAA
                      A
                     AA
Figure b9.   CRT display of cells classified as developed land in the
            demonstration area.

-------
                              -87-
                                               4444
       -cr-^  "kX
          *~«» v    4*
                                                44
Figure 50.   Overlay of major roads  and 50-ft.
            of developed land.
contours on CRT display

-------
                                     -88-
        • ooo o
ooo o    oooo o  o  o
O  OO   DO     OOOO         O
O  O» O     OOOO  O
  e  o     +0+0     o a
  e    ooooooo   o   oo  oo o  oo
 aa         OO4 440 OOOOOO     O
          O XX  **  OOOOO O   O  OO
               +444 OOOO  OO   O OO
              O444A OOOO OOOO    OO
 OO    OOl +
      O«>O4
44 4  »O +
4  4O4+4
4   44+
O 44++O
  44+444
  4+ 44
                            O
                            00
41 14  44O* OOO
 04-OODO+O
             O444444    OOO
          O  O44+44+ •  OOO
          OO  O  OA4AX*O  OO       O
           OOOOO O+AA O   OO       O
            OO 44444 O           O
           0  OO 444O O           O
              00+    O     000
              OOO  O O        O
  6   06    o   ooo  o o   ooo
0040  oo  ooooo   oao   o
040   000 00     000     OOOO
• 44+4  O   OOO    OOO
 +444  O      OO O OOO        O    O
 044    OOO   O  000      O  * XX^
OO+A      O      OO           XXX O
r    000 00  O O   OOXX         X  A O
  OOOOOO O 00 O   000 X  O       O     O
4 00   O   ++» »  004«   O.     O
          ++« * OO4>
         44|»44++«+4T4 X0*44
      OO44444A444I | | 1 4*  OO
  004 O  444.44+4+0 4OO
                             OO
               i 40     OX   44
                o o   oox  e OAA       o

   O  O 44 0  00 °t O X • O X °   4X*   3
     00 44 O +AO X  00 X  O O      •
 O 44  X        • OO            O
   4            O  OO
    00      + OOOO    O    X

   X*' °S .   4* 04     °°S  * O
4  XX     OOO       OOO   O    O
4           O*  O   O 0    O
44 444.   OOOO
44  44    O O                      OO   O
444 44 O O      O    OOO    O           »
4444444       O    O    O  X             •
4444444         O    O     IO      OO   OOl
              KEY

A - UPLAND CONIFEROUS TREES
0 - LAND SLOPE 3-8X
X - FARMS OR ESTATES
+ - CONIFEROUS TREES AND SLOPE 3-CX
V-CONIFEROUS TREES AND FARMS
D-SLOPE 3-6X AND FARMS
*-CONIFEROUS TREES, SLOPE  3-6H,,
   AND  FARMS AND ESTATES
    Figure 51.   CRT display of cells that possess various combinations of
                 (1) upland  coniferous forest, (2) a  land slope of 3 to 6%,
                 and (3) farms and estates.

-------
                              -89-
W
UJ
K
O
        1000
        800
        600
        400'
        200
-
-

"


_









1
1
VS
1
|

1
\\








1
i
^
1
^
^
i
s\;







R^






P
VN



t - UPLAND CONIFEROUS TREES
2 - LAND SLOPE OF 3-CX ^
3 - FARMS AND ESTATES


V^
vX] |\\J .
                        2     3   1-2   1-3   2-3  1-2-3


                                 CATEGORY
Figure 52.  Calculated  number of acres in the demonstration area that

            meet various  combinations of the three characteristics

            selected  for  analysis.

-------
TABLE 7.  INVENTORY OF CF.OCRAPH1C INFORMATION SYSTEM USED  BY TVA'S  DIVISION' OF  ENVIRONMENTAL PLANNING

I. Water Quality and Ecology Branch
TVA Section 26u) restrictions Periphyton data base
HvrtroloRV quality reports
Climatology reports

(1:24,000, 1:250.000)
SuitabL) ity
Quality Management
quality conditions Public

It. 316 Nonfisheries biological PRP, EPA


quality assessment
assessment OACD, PRP



USWB EXT Hap files (E)
TVA Met Data ENV PL EPA BIO STORET (E)
Perjphyton data b.ise ENV PL Technical Library (E)
Daily river bulletin W MCT Air Quality Data System (E)
qual ity reports

Corps of Engineers. USCS)
(1:24,000. 1:250,000) |
Suitability Data Q
WQF.B Index File NDRS
Stale CpoloRical Officr V MCT

Same JS above Snme as above


Same as above S.nme as above
Same ,s above Same as above





-------
TABLE 7 (continued)


and Prediction of Pollutant TPE, PRP
Transport and Effects
D. Water Quality Compliance DPP, PSSVS , FFiWD, Same .is I. A. above
OACD, EPA. Slate

MKD SV River Mow jnd elevations
Distribution and charter ist i ts
Land use patterns
Climatolojty
Land cover
rh-irai-terislics and distribution
and Control PSSVS, FFRVD. Power aqu.itic macrophytes
River flow an-i elevation
TVA fan 1 iLy locations
, data base
Mining HydroloRy
G.-OIORV
locjlions
Meteorology
Data sources

S.ine ,TH -ihovr

Aeri.il photos
daL.T-NOAA
data h.ise .
S.iroplinR station maps
TopoRraphic oaps
Project plans
Publ ished 1 itorjturc
Aerial photos
Local herbaria
TVA herbarium
Published literature
1't'f.VD Recreation Info
Aerial photos
LASDSAT tapes
USCS hydroloRic data
State mining permits
DaLa sources

Same as above

A«r ia 1 photos
dat3--NOAA
Ddi ly river bul leLin
Project plans
Publ ished 1 iterature
Aeri.il photos
LoCdl herbaria
TVA herha r ium
Dai ly river bul let in
Published literature
FF&WD Recreation Info
maps
Aerial photo*
LAXDSAT tapes
Water quality data base
TVA strip mining contracts
Supplied byb

W MGT
ENV PL
MCT

MCT
SV PL
ESV PL
LXT
V MCT
ESV PI.
FAT
W HCT
EXT
FF&VTJ
V HCT
EXT
FF&UD
ESV PL
ENV PL
W MGT
F.XT
ESV PL
EXT
PURCH
In I'ornat ion systems

Same as above
EPA STQRF.T (E)
Aquatic Biology
Information System (P)
Regional ERTS Browse File (E)
TVA Aquatic Reference System (P)
Photo Tiles (E)



Photo files (E)'
TVA herbaria (E)
Files (E)
STORET (E)
Aquatic Biology Info
System (P)
TVA Aquatic Keference System (P)
Photo files (E)
Regional ERTS
browse file (E)
TVA aerial photographic and
EPA STORET (E)
EPA BIO STORET (E)
TVA BIO INFO SYS (P)
TVA Aquatic REF SVC fp)
                                                                                                                                                                                                                                                    I
                                                                                                                                                                                                                                                   vD

-------
TABLE 7 (continued)



Coo lofty
Meteorology
Topography
Site boundaries
Hooding potential

J. Planning and Design of TAD, Public Population distribution
Municipalities Land cover
Flooding potential
Geology
Transportation networks

Hooding potential geology
Groundvater supplies
Wiisto ilist ribulioii
I] . Air Quality Branch
A. Air Quality Assessment of PRP, OACD Locations of TVA facilities
C. Spcciol Studies to Determine pheric components
l.i i») use
Monitoring Public, PHP, NDRS Regional c 1 im.i to IORV
E. Rcgionul Air Quality PSSVS Looil melcorology
Dislrihiitioii of <-n
-------
TABU 7 (continued)
Function TVA client*
III. Radiological Bygicac Branch
A. Nuclear Plant Radiological PRP
Assessment for Licensing
Approval
B. Nuclear Plant R.dlo logic* 1 P PROD
Aaaessments (or Operational
Compliance
IV. Environment Assessment and
Compliance Stair
A. Power Plant, Substation PRP
Transmission Line, and TPE
Other Major Facility Sitini OACD
Assessment
B Land Transfers and 26a P&SVS
Rev lev*
C. Outside Agency EI5 Reviews State *nJ regional
Report)
Information requirements
Water intake location!
Local meteorology
River flow and elevation
Recreation area* and use level
Hi Ik production
Site topography
Fisheries harvest
Transportation network
Residual emissions affecting
air, water and land, including chemical,
Existing land me pattern*
Proposed land use pattern*
Land ownership pattern*
(public and private)
Unit lied tract* of 1,000 or more acre*
Political and *itc boundaries
Ucllltie* location
Land, air, and water baaed trans-
Hazardous material transfer
Soil* and geologic characteriatica
and cultural area*
Projected population, indnatrial.
Residual emissions concerning air,
nolae, and aolid waate (especially
discharge data)
Existing land u«e pattern*
Propoacd land uir pattern!
Land ownership pattern* (public
ami private)
Utilities location
transportation system*
storage point*
energy, recreation demands
Site- boundaries
26a Reviews
26* Reviews
System operation data

Population distribution
data base
Daily river bulletin
Reservoir recreation
suitability data
Topographic maps (1:24,000)
Filherie* data bate
Environmental monitoring
data baae
Sampling station mapa
1:250,000 USCS maps
1:24,000 USCS mapa
TVA reservoir navigation
charts
D »tagc reservoir maps
Regional, county, local
land use plans
Aerial photographs and
remote sens tog
USDA Soil Conservation
Service*
Reservoir recreation
suitability
Population and »ociocccinomic
jectlons data base
reports
Local media and public
hearing*
Envl ronmcnta 1 1 eg la 1 ation
1:250,000 USCS maps
1:2;, 000 USCS maps
TVA reservoir navigation
charts
Regional Heritage Program
land u*e plans
sensing data
USDA Soil Conservation
Service maps
Population and •ocioeconomt c
characteristic* and pro-
jection.
report*
Same as for B, l.an
-------
TABLE 7 (continued)
                                                     TVA client3
                                                                                                                                                                                              Supplied  by_
  V.  Industrial Hygiene  Branch
      A.  Surveys  of  Potential  Employee
          Exposure to Hazardous Agents

                                               OS HA, MKD SVS
                                               PRP, OACD,  P&SVS ,  D[>P.
                                                                            Site location
                                                                            Location of noise soi

                                                                              of generated noise
                                                                                                     olume
                                                                                                       at  tim
                                                                            Site toponrnphy
Field survey data bu:

raci I ity enjtineerine
                                                                                                                                                             reports and plants
                                                                                                                                                           Topographic maps (1:24.000)
                                                                                                                                                           FaciIity maps
                                                                                                                                                           Land  use information
EN'V PI.
EXT
CLIENT
                                     F.NV PL
                                     KXT
                                     -VDRS
                                     W MCT
                                     CLIF.NT
                                     N'DRS
                                     KXT
                                     CLlEfTT
Industrial HyRi,
Map fjles (E)
Files (F.)


                Site Drawings  File (F.)
                Map  files  (E)
                Files  (E)
                Air Quality  Data  System (E)
["information  i
 Data  supplier
                                                                                                ich  iiilor(B.ition

-------
     SOURCE: vs.   FREQUENCY^ DO:VISION, FUNCTION
                  Q— FREQUENCY
                  QD— NUMBER OF FUNCTIONS
                  m — NUMBER OF DIVISIONS
             THIS MEANS 8 DIVISIONS
             USE AERIAL PHOTO8RAPHY
             A TOTAL OF 88 TIMES IN
             SUPPORT OF 28 FUNCTIONS.
                                                                                                               Ln
                                                                                                               I
FIELD  SURVEYS       FIELD       AERIAL
                INSTRUMENTATION SURVEYS
DOCUMENTS PUBLIC   PERSONAL
             FILES  COMMUNICATIONS
      Figure 53.  Sample  of the computer graphics used in the Interdivisional Study Team Report
                  on the use of geographic information by TVA.

-------
                               -96-
of the inventory and interview data.  Inventory data were first condensed
into a standardized format and coded into machine-readable form.  Statis-
tical analyses showed the TVA functions that use the most geographic
information, the divisions that are involved, and the data sources that
are most accessed.  Computer graphics software developed by this research
project was used extensively in these analyses to simplify display of
the voluminous results.  A sample of these graphics is shown in Figure 53.

     On the basis of the results of this study, a report was prepared
for TVA management with several recommendations:

1.   To define and design data bases for an expanded statistical summary
     (e.g., demographic and economic data), regional resources (geologic,
     topographic, environmental, cultural), TVA land features (legal
     boundaries), and operations impact monitoring.  These data bases
     would support TVA functions that require the largest amounts of
     geographically referenced data.

2.   To implement a series of data directories or catalogs to supplement
     current informal systems for locating geographic data.

3.   To review TVA's files and technical library systems—the agency's
     most used sources of geographic information--!© ensure that they
     are satisfying current needs.

     Systems planning with the PRIDE22 methodology is now being accom-
plished to address the first recommendation; this work is to be completed
by January 1978.   Actual implementation of suitable geographical informa-
tion data bases (pending approval by TVA's management) will proceed
during the remainder of 1978.

-------
                               -97-


                             SECTION 6

                         FUTURE DIRECTIONS

     During the next three years, many opportunities will exist to
determine the feasibility, practicality, and desirability of applying
new types of computer graphics hardware and software to environmental
analysis.

     The emphasis of the work reported here has been on the use of
storage tube graphics terminals operating through a time-sharing system
with a central computer.  A primary problem, yet unsolved, is to ensure
that a reliable, responsive computer system is available to the graphics
users when required.  TVA's computer system is overloaded, particularly
with respect to interactive users.  The maximum allowable number of
time-sharing users is only 35.  During normal working hours, the system
is used to its fullest capacity.  Even if one is successful in accessing
the system, there is always the threat of untimely system downtimes or
extremely slow response to simple commands.  These conditions are not
unique to TVA; many users of computer systems in other organizations,
both in the governmental and private sectors, report similar problems.
This presents difficulties in practical graphics implementation.  In
addition, problems such as this will grow as the demand for interactive
capabilities increase.  In many cases, history has shown that these
demands can grow almost exponentially.

     To cope with this problem, this project's research will concentrate
more on the potential use of stand-alone, "intelligent" graphics systems.
Numerous commercial systems have recently become available which offer
local graphics manipulation and data storage, FORTRAN program processing,
a relatively large CPU, and an "intelligent" terminal interface to
permit flexible communications with a large central computer.  Large
mathematical models or data storage can remain on the center computer
system.  Output data can be transferred to the smaller system for
immediate interactive graphics analysis.

     Other types of display devices will be used.  With a video (raster
scan) terminal (black-and-white or color), shading of figures can be
used to enhance the graphic effect, thus increasing communications
capability.  For example, color-shaded contour plots may be more effec-
tive in showing drift deposition over a three-dimensional representation
of uneven terrain than would an overlay of two conventional line contour
plots (one for topography and one for deposition).  Use of a stroke-writing
(refresh) terminal can produce animated sequences of dynamic behavior.
For example, a mass of complex data documenting pollutant dispersion
patterns over a period of time, perhaps in three dimensions, can be
condensed into a short 10- to 15-second "movie" sequence, which could be
shown at a public hearing.

     This project has stimulated a great deal of interest in computer
graphics in TVA.  Computer graphics is being considered as a tool to
assist in a variety of business and management applications and in
traditional engineering and scientific analysis.  TVA is considering the

-------
                               -98-
purchase of a variety of software packages that are entirely graphics-
oriented.  Using a package called DISSPLA,22 one can readily produce
pictures, maps, or two- and three-dimensional graphics with text.
Another program, called MOVIE.BYU,23 allows a user to (1) interactively
manipulate a three-dimensional figure in space, (2) accomplish calcula-
tions necessary for efficiently removing hidden lines from view, (3) cal-
culate shading patterns for color and black-and-white slides, and (4) create
individual frames of dynamic motion that can be combined into a movie
sequence.  Packages such as these should greatly enhance capability to
develop and expand practical applications for graphics.

     Finally, the demonstrations that are to be developed will emphasize
computer graphics both as a means for interaction with the computer and
as a means for communication among people.  New ways of summarizing and
displaying data will be explored.  Emphasis will be placed on graphic
communication of information concerning alternative methods of pollution
control and relationships to environmental transport and environmental
impacts.

-------
                               -99-


                             SECTION 7

                              SUMMARY

     The need for an adequate supply of electric power is becoming
increasingly evident.  At the same time, a renewed awareness of the
importance of preserving our environment has manifested itself among us
all.  To adequately meet these two requirements, engineers, scientists,
and planners must use the latest and most sophisticated methodologies
for planning energy systems and assessing impacts.  The techniques used
must provide more accurate and timely answers earlier in the planning
and design process.  Methods that permit a rapid recasting of various
analyses must be used.  A more integrated (coordinated) approach must be
taken to the assessment of alternative methods of pollutant control,
environmental impacts, and attendant costs.

     Computer graphics provides one means for significantly improving
current capabilities of environmental analysis.  Through the use of com-
puter graphics, a closer interaction between the power of the computer
and the decision-making abilities of the human mind can be achieved.
For example, an engineer can concentrate on the meaning of a particular
analysis while the computer accomplishes the necessary calculations,
manages data, and processes the results into the most meaningful form.

     Computer graphics is also an effective means for communicating
information.  Unique or previously impractical displays can be readily
generated.  New types of displays can be created and tested for effec-
tiveness.  Computer graphics can provide the means for more efficient
transfer of information among technical experts, managers, and the
public.  This, in turn, promotes improved decision making.

     Initial project research has emphasized the use of relatively
inexpensive, but versatile, storage-tube computer graphics terminals,
which depend on a large central computer for computations and manipu-
lation of graphics display.  Specific project activities accomplished
during the first two years of work include (1) stimulating interest in
the use of computer graphics and identifying potential applications,
(2) acquiring suitable graphics hardware capability that is accessible
to potential users, (3) developing general-purpose software that promotes
the use of graphics, (4) identifying and solving difficulties in implemen-
tation that potentially limit the use of graphics, and (5) developing
demonstrations to illustrate benefits to be derived from using computer
graphics for environmental analyses.

     Demonstrations addressing various difficulties have been developed
to facilitate data display, interactive analysis, and mathematical
modeling.  Graphics has been successfully applied to analyses of air and
water quality, radiological hygiene, industrial hygiene, analysis of
socioeconomic impact, and spatial analysis by means of geographically
referenced data for siting facilities.  Although developed as demon-
strations, most techniques have been adopted for routine use.  The
result has been a savings in time and cost for certain analyses.  Also,
graphics displays previously considered impractical are being readily
accomplished.

-------
                               -100-
     Several types of computer graphics hardware other than storage-tube
devices show promise in benefiting environmental analysis.  For example,
devices with capabilities for color and animation might be used effec-
tively to describe complex environmental phenomena.  Powerful graphics
software is rapidly becoming commercially available which will permit
rapid development of complex displays that are highly tailored to a
specific need.  One yet unresolved problem is that of providing a reliable,
responsive graphics capability to users who must depend on a conventional
time-sharing system through a large central computer.  The use of stand-
alone or intelligent graphic systems that can communicate with larger
computers may provide a means for solving this remaining difficulty.

-------
                               -101-


                            REFERENCES

 1.   Tektronix,  Inc., P.O.  Box 500,  Beaverton,  Ore.   97077.

 2.   California  Computer Products,  Inc.,  305  North Muller  Street,
     Anaheim, Calif.  92803.

 3.   Gnanadeskan, R.  Methods for Statistical Data Analysis  of Multi-
     variate Observations.   John Wiley & Sons,  Inc.,  New York, 1977.

 4.   Anderson, E.  A Semigraphical  Method for the Analysis of Complex
     Problems.  Technometrics, 2(3): 387-391, 1960.

 5.   Chernoff, H. and Rizvi, M. H.   "Effect on  Classification Error  of
     Random Permutations of Features in Representing  Multivariate  Data
     by Faces.  J. Am. Stat. Assoc., 70(351): 548-554,  1975.

 6.   Morris, M.  F.  Digital Computer Usage.  In McGraw  Hill  Encylopedia
     of Science and Technology, 1975. pp. 156-161.

 7.   Morris, M.  F.  Kiviat Graphics and Single  Figure Measures Evolving.
     Computerworld, Feb. 9: 17-18,  1976.

 8.   Siegel, J.  H. , Farrell, E. J.,  Goldwyn,  R. M.,  and Friedman,  H. P.
     The Surgical Implications of Physiologic Patterns  in  Myocardial
     Infarction Shock.  Surgery, 72(1): 126-141, 1972.

 9.   Goldwyn, R. M., Farell, E. J.,  Friedman, H. P.,  Miller, M., and
     Siegel, J.  H.  Identifying and Understanding Patterns and Processes
     in Human Shock and Trauma.  IBM J. Res.  Dev., May: 230-238, 1973.

10.   Brock, E. S. and Yake, W.  A Modification  of Maucha's Ironic  Diagram
     to Include Ionic Concentrations.  Limnol.  Oceanogr.,  14(3): 933-935,
     1969.

11.   Thornton, K. W. and Lessen, A.  S.  Sensitivity  Analysis of  the  Water
     Quality For River-Reservoir Systems Model.  Waterways Experiment
     Station Miscellaneous Paper V-76-4,  U.S. Army Corps of  Engineers,
     September 1976.

12.   Jones, H. C., Lacasse, N. L.,  Liggett, W.  S., and  Weatherford,
     Frances.  Experimental Air Exclusion System for  Field Studies of
     SO  Effects on Crop Productivity.  E-EP/77-5, Tennessee Valley
     Authority,  Division of Environmental Planning,  December 1977.

13.   Slawson, P. R.  Vapour Plume Theory and Example  Computer Programs.
     Prepared for the Tennessee Valley Authority, Division of Environmental
     Planning, Air Quality Branch,  Muscle Shoals, Ala., by Envirodyne Ltd.,
     Waterloo, Ontario, Canada, TVA Contract TV38636A,  January 1976.

-------
                               -102-
14.  Slawson, P. R. and McCormick, W. J.  Drift, Vapour and Dry Plume
     Model with Documentation.  Prepared for the Tennessee Valley
     Authority, Division of Environmental Planning, Chattanooga, Tenn.,
     by Envirodyne, Ltd., Waterloo, Ontario, Canada, TVA Contract TVA33332A,
     June 1976.

15.  DeAngelis, D. L. and Tharp, M. L.   A Computer Code For the Three-
     Dimensional Graphing of Multiple Surfaces-Applications to Ecology.

16.  Gaume, A. N. , Brandes,  R. J., and Duke, J.  H., Jr.  Computer Program
     Documentation for the Reservoir Ecologic Model TVAECO, with Tims
     Ford Reservoir Simulation Results.   Prepared for the Tennessee
     Valley Authority by Water Resources Engineers, Austin, Tex.,
     February 1975.

17.  Babb, M. C. and Bruggink, D. J.  Deep Reservoir Temperature Model,
     Tims Ford Reservoir.  Special Projects Staff Report, Water Quality
     and Ecology Branch, Division of Environmental Planning,  Tennessee
     Valley Authority, Chattanooga, Tenn., April 1975.

18.  Gordon, J. A.  and Babb, M. C.  Problems Associated with the Validation
     and Use of Reservoir Water Quality Models.   In Proceedings of the
     5th Annual Environmental Engineering and Science Conference, University
     of Louisville, Louisville, Ky., March 3-4,  1975.

19.  Howard, E. E., Rowland, E. B., and Smart,  C.  W.  Review of Power
     Facility Siting Methodologies with Emphasis on the Application of
     Computer-Assisted Geographic Information Systems.   Tennessee Valley
     Authority, Division of  Forestry, Fisheries  and Wildlife Development,
     Norris, Tenn., August 1976.

20.  Smart, C. W., Rowland,  E. B., and Baxter,  F.  P.  Conceptual Application
     of a Computer-Assisted  Geographic Information System to Address
     Generic Power Plant Siting Objectives, Tennessee Valley Authority,
     Division of Forestry, Fisheries and Wildlife Development, Norris,
     Tenn., January 1977.

21.  Laboratory for Computer Graphics and Spatial Analysis, Graduate
     School of Design, Harvard University, Cambridge,  Mass.  02138.

22.  M. Bryce and Associates, Inc.  Profitable Information by Design
     Through Phased Planning and Control, Cincinnati,  Ohio, 1976.

23.  Integrated Software Systems Corporation, 4186 Sorrento Valley
     Blvd., Suite N, San Diego, Calif.   92121.

-------
APPENDICES

-------
                                 A-l

                            APPENDIX A

   SUMMARY OF POTENTIAL ENVIRONMENTAL ANALYSES TO WHICH COMPUTER
                     GRAPHICS COULD BE APPLIED


     Various environmental analyses could potentially be used as the basis
for demonstrating the utility of computer graphics.   The inventory described
in the appendix was conducted by (1) directly contacting potential users,
(2) drawing on the experience of this project researcher, and (3) distribut-
ing a questionnaire to branch and staff personnel of TVA's Division of
Environmental Planning.

  I.  Air Quality Analysis

      A.  Modeling of the downwind distribution of air pollutants from a
          given arrangement of mobile point sources
      B.  Analysis and display of meteorological data
      C.  Formulation and cost modeling of air quality monitoring network
      D.  Analysis and display of air quality data obtained from field
          measurements to determine data trends and compliance with
          regulations
      E.  Display of predictions of modeling concerning transport and
          transformation of chemical species of air pollutants
      F.  Display of results of model calculations showing the behavior of
          emissions from pollution  control devices (e.g., stacks, cooling
          towers, scrubbers)
 II.  Water Quality Analysis

      A.  Site screening potential analysis to determine impacts from a
          specific facility site and a particular project design
      B.  Display and  analysis of water quality trends
      C.  Predictions  of the transport and effects of water pollution
          spills
      D.  Display of predictions of natural physical, chemical, and bio-
          logical phenomena in aquatic systems through the use of
          mathematical models
      E.  Analysis and modeling of non-point sources of pollution


III.  Radiological Hygiene Analysis

      A.  Development  of radiological dose model
      B.  Determinations of nuclear plant layout
      C.  Cost-benefit analysis of systems for controlling radiological
          waste
      D.  Real-time accident analysis of contaminant dispersion and
          transport
      E.  Radiation shield and analysis
      F.  Cost-benefit analysis of tritium waste management program
      G.  Statistical  analysis of environmental monitoring data

-------
                              A-2


IV.   Safety and Industrial Hygiene

     A.   Display of accident occurrence data
     B.   Analysis and display of field measurements  of potentially
         hazardous agents in the working environment
     C.   Modeling of noise control  in working and community environments
 V.   Generalized Data Reduction,  Analysis,  and Display

     A.   Development of an interactive  time series  analysis  package
         with graphics displays
     B.   Development of interactive  graphics for  presenting  the  results
         of exploratory statistical  analyses
     C.   Development of a  general-purpose  contouring  routine
     D.   Development of a  three-dimensional perspective plotting routine
         for environmental data
     E.   Development of off-line  digitizing routines  and  routines for
         on-line accessing of  the flexible-disc memory units

-------
                                 B-l

                            APPENDIX B

         DIFFICULTIES WITH COMPUTER HARDWARE AND SOFTWARE
          ENCOUNTERED WHEN IMPLEMENTING DEMONSTRATIONS OF
                         COMPUTER GRAPHICS


      The following difficulties with hardware and software were addressed
by this project between September 1975 and December 1976.  The solution
of these problems has allowed demonstrations of computer graphics to be
developed for a much wider range of applications than was feasible pre-
viously.

 1.   Problem:  Complex, high-density information display could not be
      generated in a reasonable amount of time because of the 300-baud
      telecommunications rate.

      Solution:  Provided necessary justification for upgrading TVA's
      telecommunications system to 1200 baud and the purchase of the
      high-speed models.

 2.   Problem:  Effective, interactive engineering analysis was precluded
      as a result of extremely slow system response.

      Solution:  Arrangements were made to obtain time-sharing services
      from Computer Sciences Corporation (CSC), INFONET, Los Angeles,
      California.  Learning a new time-sharing command language and the
      use of the unique features of this system was required.

 3.   Problem:  To minimize on-line computer time, a scheme was needed
      whereby data or programs could be entered and stored off-line and
      then put on TVA's systems at a TSO session.

      Solution:  A documented scheme was developed to enter information
      manually off-line and store it on the flexible-disc memory.  On-line
      software could then be executed to store the data and program in a
      user's on-line TSO library.

 4.   Problem:  The capability for digitizing maps and graphs off-line and
      entering that information into the computer during an interactive
      session was needed.

      Solution-:  Procedures for activating the graphics tablet and flexible
      disc were established.  On-line software was written to convert coded
      bit information stored on the flexible disc to user coordinates.  This
      information could then be stored on the flexible disc or a TSO data
      set, edited, or translated into moves and draws on the screen.

 5.   Problem:  Many graphics software packages available on TVA's computer
      system, such as three-dimensional, contour, and spatial analysis pack-
      ages, are oriented for batch processing.

      Solution:  Interactive analysis routines were written to call the
      basic graphics software packages.  Simplified interactive routines
      vere developed by means of the basic algorithms of the original
      software with TEKTRONIX graphics subroutines.

-------
                                 B-2


 6.   Problem:  Several widely used TEKTRONIX graphics output routines
      handle input information in ASCII decimal equivalent (ADE) form.

      Solution:  A general subroutine was written to convert ADE to
      alphanumeric characters (A/n or A/N) to ADE.

 7.   Problem:  To minimize array sizes in interactive analysis, only
      that particular set of data values needed for an analysis should
      be read into a program.  To accomplish this effectively,  the data
      must be stored in direct,  random-access data  sets.   The method for
      generating such data sets  on TSO is not readily apparent from the
      TSO user's manuals.

      Solution:  A procedure for establishing this  type of data set was
      established.  A command language program was  developed to make the
      necessary allocations and  call a system's utility formatting program.

 8.   Problem:  To effectively use time-sharing systems other than TVA's,
      a method was needed to convert programs residing on TVA's IBM 370/
      165 and enter them on the  other systems.

      Solution:  Procedures were established for formatting programs prop-
      erly on the IBM 370/165 and storing them on the flexible-disc memory
      unit operating in the paper tape mode.  Procedures for transferring
      this information to and from CSC using their  DATA command and system
      utility programs were then developed.

 9.   Problem:  Numerous "bugs"  were discovered in  the TEKTRONIX software.

      Solution:  These problems  were identified and documented.  System
      analysis representatives,  advised of the problems,  have corrected
      these errors.

10.   Problem:  Complex analysis techniques that involve access to several
      files, calling various programs, and allocating storage require that
      the person running the analysis know TSO command procedures and enter
      it properly.

      Solution:  Command language programming techniques have been developed
      to simplify command procedure processing.  The user simply answers
      yes-and-no questions to execute the various programs or system tasks.

11.   Problem:  Achieving nonstandard graphics displays through the use of
      Plot-10 software package.

      Solution:  Plot-10 software has numerous software "hooks" for incor-
      porating user-written plot routines and accessing internal variables
      that are useful in creating specialized plots.  Routines that explore
      the use of these features  have been developed.

-------
                                 B-3

12.    Problem:  Outside time-sharing services can only be readily accessed
      through the Federal Telecommunication System (FTS) phones; this ties
      up FTS phone lines and a TVA switchboard line.

      Solution:  Under certain conditions, the General Services Administra-
      tion (GSA) will grant access to "800" numbers that are linked to the
      outside time-sharing services.  The necessary forms were submitted to
      GSA for approval of eight TVA users.

-------
                              C-l
                            APPENDIX C

       TVA SOCIOECONOMIC IMPACT ASSESSMENT METHODS PROJECT-
               DEVELOPMENT OF SITE SCREENING METHODS

                           INTRODUCTION

     The development of a methodology for screening potential sites for
power plants according to socioeconomic criteria has been a major aspect
of the second year of work in TVA's Socioeconomic Impact Assessment
Methods Project.  There have been two steps to the development of the
method:  (1) to compile data for a particular area on the indicators
developed during the first year of work and (2) to develop the computer-
assisted method for analysis and display.

     During the first year of the project, a set of indicators to be used
on a county-level basis was developed to provide the data framework for
site screening.  The second year of work was designed to test this set of
indicators, amend them if necessary, and develop a method for analyzing
and displaying the overall suitability of different counties for absorb-
ing the impact of a large power generating facility.

                          DATA COLLECTION

SELECTION OF THE TEST SITE

     Twenty-two counties in East Tennessee (Figure C.I) were selected for
testing the site screening method for several reasons.  The TVA Regional
Planning Staff had just completed a site screening exercise by means of
existing agency methods.  Therefore, repeating this exercise seemed to be
a good test of whether a computerized system with fairly extensive data
needs would show different results from those of the more intuitive, tradi-
tional planning method.  Also, the proximity of these counties to Knoxville
made data collection easier and less time-consuming.  The area is also a
good test site because it contains within it a wide variety of counties,
ranging from highly urbanized Hamilton County to Hancock County, a small
mountain county that has long been one of the poorest counties in the
country.

     Initially, Knox and Sullivan Counties were not included in the study
area.  Knox County was excluded because it was felt to be too highly
developed along the main river channel to be a reasonable site for addi-
tional power facilities.  Sullivan County was eliminated because it was
felt that water supplies that far upriver would not be sufficient to make
a power plant feasible.  However, both Knox and Sullivan Counties are now
being added to the study area.  On reconsideration, it was felt that the
reasons for excluding them are only two of many reasons that may eventually
exclude any one of these counties from further consideration as power plant
sites.  One important reason for including them is that any power plant will
have impact on the counties adjacent to the one in which it is located, not
only the county in which it is located.  Therefore, the same types of socio-
economic data are needed for counties that may be close to power plant sites.
At this time, data for these two counties have been added, but analysis has
yet to be carried out on a 24-county basis.

-------
,COOKEVILLE
                                                                                                          n
                                                                                                          N3
Figure C.I.  Regional location map for site screening area  in East  Tennessee.

-------
                              C-3
COLLECTION OF DATA

     Data were collected during July, August, and September of 1976.
Many of the data were collected from published sources, although there
were several instances for which a trip to Nashville, Chattanooga, or
Johnson City was necessary to supply missing information.  One diffi-
culty in data collection was the fact that the development districts
located in the study area are not always consistent in their method of
recording data.  Types of the data that were published and up-to-date
in one development district were not even collected in another.

     Other data that were sought were kept by the various State agencies
on a county and municipality level.  To get the most up-to-date informa-
tion, some personal interviews and two trips were necessary.  There is a
time lag of a year or more in publishing many of the data before they
are readily available.

REVISION OF INDICATORS

     The indicators were modified, new ones were added, and some were
eliminated at several points throughout development of the site screening
system.  During data collection, several indicators were found to be  mean-
ingless or too difficult to obtain; others were changed or further detailed
from what was specified in the indicator system outlined after the first
year of work.  The changes that were made will be described separately for
each indicator listed.
Highway, Sewer, and Water Systems

     This indicator was eliminated at the end of the data collection period.
An attempt was made to visually screen out those areas in which no primary
highway was located within 5 miles of a major waterway, thus eliminating
counties that had no eligible areas by this criterion.  However, no counties
could be eliminated by this procedure.

     Several other access measures were considered.  Working with mapped
data when all other indicators were on a gross county level proved diffi-
cult, therefore, a county-level access indicator was sought, but none
proved satisfactory.  One reason for this is that the major transporta-
tion routes have tended to develop along rivers, which also define the
likely locations for power plants.  Therefore, any simple measure was
ineffective in screening counties because all looked equally good.  The
final decision was to rely on the engineering studies by other divisions
within TVA to eliminate any inaccessible sites.

     Information on all the major and minor highways in the study area was
collected from the Tennessee Department of Transportation (DOT), and
details on the condition of each of these roads were obtained from knowl-
edgeable staff members of DOT wherever possible.  Collection of this data
was rather difficult, requiring a trip to regional offices in Chattanooga
and personal interviews with highway staff members.

-------
                              C-4
 The  sewer  and water systems could not be located on maps because this infor-
 mation was not  readily available.  Although boundaries for sewer and water
 utility  districts were available for some parts of the region, they were
 not  available for all counties.  Also, utility districts seldom operate
 sewers throughout their entire service boundaries.  Most of these districts-
 both water and  sewer—are small and do not keep good up-to-date maps of
 their systems.  Information from these districts could have been obtained
 if we had been  able to spend the effort to transfer raw data to usable
 mapped form.  The problem would then remain of having mapped sewer and
 water data that were not necessarily on a county basis, would not be
 compatible with the other indicators, and could not be readily computerized.
 To get a measure of the water and sewer service available in various
 counties, an indicator was substituted which is the percentage of dwell-
 ing  units served by public water and sewer systems.  This is included in
 the  public facilities indicators described below.

 Judgmental Ranking or Rating Made of Each of the Public
 Facilities Systems and Displayed Graphically by County

     Several aspects of public facilities were considered—recreation,
 health, police and fire service, and education.  Indicators were developed
 in each of these public service areas.

 Recreation—

     The indicator developed for recreation capacity was the recreation
 acres per capita in a county.   This was the total acreage in parks in
 recreation facilities, including local, county, state, and national park
 systems.  Information was sought on recreation expenditures by public
 bodies by each county, but data were not readily available.  Examination
 of these figures showed that the figures were distorted because of the
 presence of large national parks and forests in several counties.  While
 these parks are, in fact, a recreation resource, they are not of the
 same nature as are local ballfields or other neighborhood or municipal
 facilities.  Therefore, an alternative indicator was developed, which is
 the  local recreation acres per capita.  This eliminates State and Federally
 owned recreation lands from the calculations.

 Health-

     Several indicators were considered as measures of health care.
 After discussions with representative staff members of TVA's Health
 Resources Staff, it was decided that the best single indicator of health
 care was the availability of physicians.  The indicator developed was
 the number of persons per physician (including osteopaths) in each
 county.  Other indicators were rejected for various reasons.  The number
 of hospital beds per capita was not used because hospital services are
primarily a regional phenomenon, and one county need not have a great
deal of hospital space if it is close to a county with available hospital
beds.  Also, there are generally enough hospital beds in this region.
The major problem with health care is the lack of primary personnel,
 such as doctors, in the more isolated rural areas.  Our data did show
that the smaller, more rural counties have a much greater population
per doctor than do the urban counties.  Meigs County had no doctors at
all.

-------
                              C-5
Police--

     The indicator developed for police services was the expenditures
per capita for police departments by cities and counties.  To convert
the collected data to a county level, a weighted average was developed
for each county to consider the expenditures of municipal police depart-
ments and county sheriffs' departments.  Expenditures in 1975 were
obtained for every police department and sheriff's department in the
study area, and per capita expenditures were calculated for each.  The
population served by each police force was then used to weight those
expenditures.  It is assumed that sheriff's departments serve only
outside of municipalities.  This weighted average was calculated as
shown in the following spot table.
Police
force
Expenditures
per capita
(a)
% of total
county
population
(b)
Weighted
expenditures
(a x b)
Total
county
weighted
average
Bradley Co.
$ 4.10
0.612
$ 2.51
$13.75
Cleveland
City
Charleston
City
29.53
14.08
0.374
0.014
11.04
0.20
Thus, the total  county weighted  average of $13.75 reflects the fact that
about 38% of  the county's population  lives in cities, where there are
considerably  higher  levels  of  expenditure than in the county-served areas,
where 61% of  the people  live and where the sheriff's department spends
only about  $4.00 per capita on police protection.

Fire Protection--

     A  satisfactory  fire protection indicator has not been developed.
Data was collected on fire  department expenditures per capita, but this
indicator has been rejected because of the problem of collecting reliable
data.   Expenditure figures  are available for the cities, and the only pro-
blems are occasional incomparability  of data (for example, where rescue
squads  are  included  within  the fire department in one city and separately
in another  city) and the lack  of centralized data sources.

     The major problem arises  in the  rural areas that are served by
private or  volunteer fire departments.  These small fire departments
keep very poor records,  if  any.  They also may perform nonfire services,
such as rescue or ambulance services. Also, the expenditures are not
comparable, even if  the  correct  figures are known.  In some instances,

-------
                              C-6
 firemen  are paid only  for the fires that they actually fight; in other
 instances, as  in a  city  fire department, they are on a salary.  In still
 other  cases, volunteers  are used, and only one or two paid firemen may
 be  employed.   These rural fire departments are also difficult to locate,
 and it is unlikely  that  reliable information could be obtained from them
 regularly.

     As  an alternative indicator, the use of ratings from State insurance
 fire departments is being investigated.  City and rural fire departments
 are rated differently, and much of the rating depends on such things as
 water  supply and pressure rather than on the proficiency of the fire
 department, its level of expenditures, or its equipment.  However, it is
 felt that this may be a  suitable indicator and the best one available
 across the State for fire protection.  These data are being analyzed.

 Sewer  and Water Facilities—

     As  indicated above, a new indicator has been added—the percentage
 of  county dwelling units served by public sewers and water systems.  This
 figure,  derived from the 1970 Census, is somewhat limited in its useful-
 ness because it is several years old.  However, on the basis of our
 investigations, this figure is the only readily obtainable measure of
 sewer  and water facilities for the purposes of site screening.

 Education—

     The education indicators are directed only at the local public school
 system because the local public schools will feel most of the impact of the
 added population from power plant construction.  For purposes of site
 screening, higher education, vocational education, and private schools are
 not considered.

     Several education indicators, most of which are readily obtainable
 from the State Department of Education, were considered.  Four indicators
 were finally included which represent some of the more important dimen-
 sions  of public education quality:  (1) total expenditure per pupil, (2)
 percentage of  students who are in overcrowded classrooms, (3) average
 teacher  salary, and (4) number of pupils per teacher.  For each indicator,
 a weighted average was calculated for each county in a manner similar to
 that done for police expenditures.  However, instead of using the propor-
 tion of  total population in the various jurisdictions, the total enrollment
 of  each  school system within each county was used to weight the statistics
 to  come up with a county average.

     Expenditures per pupil, teacher salary, and pupil/teacher ratio are
 all taken directly from the Annual Statistical Report of the State Depart-
ment of Education.  The percentage of students in overcrowded conditions
was calculated from the State Department of Education's Report on Waivers.
Any time that a classroom has more than the allowable number of students
 (25 to 35, depending on the grade), the school system must obtain a waiver
 from the State.  From the data in the Annual Report on Waivers, the per-
 centage of total students who are in classes with more than the allowable
number of pupils was calculated.

-------
                              C-7
Existence of a Planning Commission and a Comprehensive Plan;
and Existence of Zoning and Subdivision Regulations

     The quality of planning and public administration within each county
was characterized according to the existence of planning commissions,  com-
prehensive plans, and zoning and subdivision regulations within that county.
A list was prepared for each of these planning and administration tools,
indicating which cities and counties had planning commissions either in
existence or active, which had comprehensive plans, which of these plans
were current or not, and which had regulations on zoning and subdivision.

     Some manipulation of the data was necessary because the information
had to be on a county level.  In some counties, municipalities may have
zoning, but the areas outside the municipalities do not.  Similarly, there
may be subdivision regulations in effect in portions of the county and not
in others.  In some cases, there is a county-wide planning commission; in
others, planning commissions exist only within the municipality.  Therefore,
for each of the four planning and public administration tools, a new
statistic was calculated which is the percentage of the population in that
county that resides in a jurisdiction covered by a planning commission, a
comprehensive plan, a zoning ordinance, or subdivision regulations.  For
example, in Anderson County, 100% of the population is covered by a planning
commission, zoning, and subdivision regulations.  However, only 65% of the
population lives within an area that has a comprehensive plan.

     These calculations allow one to differentiate between counties with
county-wide plans and subdivision regulations and those in which only the
municipalities or only some of the municipalities have such tools.  Also,
some counties may have zoning ordinances that do not apply within munici-
palities, and the municipalities may not have zoning ordinances.  Thus, this
indicator provides a gross measure of the commitment of that county as a
whole to planning and public administration as measured by these four tools.
In making these calculations, localities that had a comprehensive plan that
was out of date or a planning commission that only existed on paper were
not included.

     A third indicator, the existence of a budget process including capital
budgeting was dropped because analysis of the data indicates that capital
budgeting is not a useful indicator.  Our information suggests that there
is no common definition of the capital budgeting process.  The fact that
many communities with meager administrative resources report the use of a
capital budgeting process casts doubt on the validity of this indicator as
a measure of administrative capacity.

Percent of County Population Living in Urban Places

     This indicator was obtained directly from the 1970 Census.

     Urbanization is a very important criterion in avoiding social impacts;
experience has shown that small towns suffer the most significant negative
impacts from the construction of power facilities.  Therefore, it was
attempted to develop a more sensitive indicator of urbanization.

-------
                              C-8
     As an alternative indicator, a visual screening was proposed that
would eliminate counties in which no area is within 10 miles of a city
of 10,000 population or more.  This criterion eliminated three counties
from consideration:  Rhea, Hancock, and Claiborne.  However, elimination
of counties was not considered realistic at this site screening stage
because nothing in our studies or TVA's experience suggested that any
single socioeconomic factor is essential.  Therefore, another proposal
was made to merely downgrade those three counties in the rating system
that was to be developed.  However, this indicator was finally discarded
because it would have been the only one requiring a visual screening and
thus could not have been incorporated into a computerized screening
system.

     The percent of population living in urban areas thus remains the
only indicator of urbanization.

Labor Force Availability to be Characterized

     TVA's Human Resources Staff regularly compiles information that can
be used to show labor force availability.  Although a gross measure could
be obtained from the 1970 Census, a much more accurate and up-to-date
picture is available within TVA.   This data has not yet been made avail-
able to us, but TVA's Regional Planning Staff is collecting the data.

Vacancy Rate of Housing for Sale and for Rent

     This indicator was collected from the 1970 Census, because more up-
to-date information is not available.

Percentage of Housing Units that Were Built Before 1950

     This indicator was also collected from the 1970 Census.

Population Change from 1960 to 1970 and Projected
Population Change

     The population change from I960 to 1970 for each county was obtained
from the 1970 Census.  The projected population growth for the period
1970 to 1990 was obtained from the U.S. Department of Commerce, Bureau
of Economic Analysis.  These figures are already collected within TVA's
Division of Navigation Development and Regional Studies (ND&RS).

Unemployment Rate

     This indicator was obtained for each county from the most current
data of the State Department of Employment Security.

Median Family Income

     The 1970 Census was used for this indicator.

-------
                              C-9
Percentage of Families and Individuals Receiving
Welfare Payments

     The most recent data that could be obtained for this indicator is
from the 1975 report of the State Department of Human Services.

Proportion of the Population in the 20- to
44-Year Age Group

     Information for this indicator was obtained from the 1970 Census.

Percentage of Population Living in Same House
Five Years Ago

     This indicator was part of the system during the early phases of the
study.  The percent who have moved in the last five years was intended to
be a measure of social stability, particularly out-migration.  Further
analysis, however, showed that this is not a good indicator of out-migration.
Clearly, anyone who had migrated out of the county would not appear in this
statistic.  Therefore, it shows geographic mobility within counties and in-
migration.  This was evident from the results showing that some of the most
urban  counties had a high score on this indicator.

     A list of each of the  indicators entered into the  computerized data
sets is shown in Table C.I.

                DEVELOPMENT OF DESCRIPTIVE INDICES

PURPOSE

     The next step was  to  develop  some  indices,  or  combinations of  indi-
cators, that  could be  used  to  describe  the counties'  relative suitability
as power plant  sites.  Each index  is  intended to describe  a  functional
system of  community  life,  so  that  together they  will  provide  an accurate
representation  of  the  socioeconomic  characteristics of  the county as  a
whole. It  is  important  that  the  indices  be  small in  number,  yet  compre-
hend all potential  areas  of socioeconomic impact.   Furthermore, the indices
are not necessarily equal  in importance.   Some  functional  areas may be more
vulnerable  to  socioeconomic impacts  than  others, and  impacts  in some  func-
tional areas  affect  local  residents  more  than  they  do in others.  Therefore,
in computing the overall suitability of counties, the indices must be
weighted  in some way to account for  these differences in importance.

     The  indices relate to the considerations  in site screening developed
during the first year of work; however, they may not  be exactly the same.
The purpose of developing indices,  in summary,  is two-fold:   (1)  to reduce
a large number of indicators to a manageable size and (2)  to provide  a
descriptive model of each county's socioeconomic conditions  and potential
for absorbing or benefiting from such impacts.

-------
                                   C-10
                  TABLE C.I.  SITE SCREENING INDICATORS

Stat.
programs
number
1
2
3

4
5
6
7
8

9

10

11

13
14
15
16
17

18
19
20
21
23
24
Graphics
programs
number
10
11
12

14
15
16
17
18

19

20

21

23
25
26
27
28

29
30
31
32
34
35
Indicator
Recreation: recreation acres per capita
Health: local physicians per thousand persons
Police: expenditures per capita (weighted
average of each county with its cities)
Expenditures per pupil
Percent students overcrowded
Average teacher salary
Number of pupils per teacher
Percent of county population in jurisdiction
having a planning commission
Percent of county population in jurisdiction
having a comprehensive plan
Percent of county population in jurisdiction
having a zoning ordinance
Percent of county population in jurisdiction
having subdivision regulations
Percent of county population in urban places
Housing vacancy rate (1970)
Percent of housing built before 1950
Percent of change in population (1960-1970)
Percent of change in population projected
(1970-1990)
Unemployment rate
Median family income (1970)
Percent of population receiving welfare
Percent of population in 20- to 44-year age group
Percent of population served by public water
Percent of population served by public sewer

Separate code numbers were used for the statistical package and the
graphics package.

-------
                              C-ll


ANALYSIS OF THE INDICATORS

     The first step in developing indices was to analyze the indicators.
Examination of the distributions of actual statistics collected on each
indicator provided a sense of the information and its quality and use-
fulness.  Also, statistical tests were run to identify any redundant
indicators that might be eliminated to streamline the data system.

Methods

     The methods used for analysis included visual examination of the
distributions, Pearson and Spearman rank-order correlations on all pairs
of indicators, and scatter diagrams of several pairs of indicators.

Findings

From Visual Examination of Distributions—

     Recreation—It was noted that the distribution of recreation acreage
was skewed due to the presence of large tracts of State and Federal park
lands in some counties.  Because this was felt to be an undesirable dis-
tortion of the level of recreation resources, the alternative indicator
(deleting State and Federal lands) was developed as described previously.

     Education—The range of values for all the education indicators was
fairly small.  This is because, unlike other public facilities and
services, the public schools are governed by State minimum standards in
many aspects of their operation.  The "percent in overcrowded classrooms"
indicator showed a much different distribution—including rank order—
than did the other three education indicators, which tended to be similar.
Counties with very few of their pupils in crowded conditions included
both wealthy counties, which had more than adequate facilities, and poor
counties, in which, although facilities may not be up to par, continued
out-migration has caused declining school enrollments.

     Planning and public administration—These indicators were manipulated
into county level percentages, and many counties clustered at the maximum
potential score (100%).  As expected, the urbanized areas tend to use more
of the planning and administration tools and therefore have higher scores
on these indicators.  This is probably a good indication that they will
be better able to respond to the demands brought by rapid growth  resulting
from a power plant.

     Because the capital budgeting indicator is of doubtful validity as a
measure of administrative capacity, it will be eliminated from the data
set.

     Indicators of housing, unemployment, welfare, and population age
structure all had distributions that tended to cluster within a relatively
narrow range of values.  The knowledge that these and other distributions
had many tied or closely grouped scores contributed greatly to the decision
to limit the use of ranks in comparing the counties.

-------
                              C-12
Correlations—

Two types of correlations were run on the indicators:  (1) Pearson product-
moment (r) using the raw scores and (2) Spearman rank-order (rho) using the
rank orders of the counties.  The findings of these statistical tests were
not as revealing as had been hoped.  The principal purpose of these tests
was to find indicators that were very highly correlated in order to identify
possibilities for elimination or combination of indicators.  The only two
indicators that were identified as possibly redundant were pupil expendi-
tures and average teacher salaries—the rank-order correlation of these was
0.8 (out of a possible 1.0).  However, because this was the only possible
redundancy that could be identified both statistically and intuitively, it
was decided not to eliminate any indicators.

     Two unexpected correlation coefficients are worth noting—education and
planning and administration.

     Education—The percentage of pupils overcrowded was not highly corre-
lated with the other education indicators.  This lack of correlation
resulted from this indicator's peculiar distribution.  As noted before,
the school systems without much crowding tended to be either relatively
wealthy and urbanized or poor and rural with a history of out-migration.
The other indicators tended to show the urban, wealthy counties with con-
sistently better scores than the poor, rural counties.  Thus, the over-
crowding indicator seems to be measuring a very different dimension of
public education conditions than are the other three indicators—perhaps
the immediate physical capacity rather than overall quality or growth
adsorption capacity.

     Planning and administration—The planning and administration indicators
were not all highly correlated with each other or with urbanization, as was
expected.  The existence of subdivision regulations was not correlated with
either capital budgeting (r = 0.08) or urbanization (r = 0.17), nor was the
existence of planning commissions correlated with urbanization (r = 0.17).
This may say something about the way in which the indicators were developed.
For instance, a county may have a high score on urbanization either because
it includes a major city or because most of its population lives in small
towns that just barely meet the Census "urban place" definition.  These
small towns may not have subdivision regulations or capital budgeting; and,
as noted before, the capital budgeting indicator has serious flaws as a
measure of public administration capacity and will be eliminated.

Scatter Diagrams—

     These diagrams show graphically the relationship between two indicators.
Scatter diagrams were plotted for several pairs of indicators, using the
rank orders of counties on each.  The diagrams were developed only for those
pairs whose correlations were different than expected or in some way note-
worthy.  Examination helps to identify the reasons for the statistical
relationship, particularly where there is no linear statistical relation-
ship between two indicators intuitively felt to be associated.  For example,
the scatter diagram of "percentage of urban places" with "percentage with
planning commissions" shows a definite positive relationship at the lower

-------
                              C-13
end of the distribution of planning commission indicator,  although the
correlation coefficient is only 0.17 (almost no relationship).   However,
the diagram clearly shows that the reason for the low correlation is  the
very large number of counties (of all degrees of urbanization)  for which
100% of the county is covered by a planning commission.

     Thus, the scatter diagrams helped to confirm some intuitive relation-
ships and provide a better understanding of why some unexpected correlation
coefficients were obtained.  This exercise did not, however,  add a great
deal to the analysis of the indicators, but served mainly as  a  check  to
verify judgmental interpretations of the data.

               DEVELOPMENT OF THE COMPUTER-ASSISTED
                    ANALYTIC AND DISPLAY METHOD

PURPOSE

     A major purpose of the site screening phase of the socioeconomic
impact assessment project was to demonstrate the use of computer graphics
in this planning effort.  Funds from the Environmental Protection Agency
pass-through grant were used to purchase a Tektronix cathode ray tube
(CRT) computer terminal, with a floppy-disc data storage component and
hard-copy attachment.  Thus, the analytic and display method was designed
to be compatible with the Tektronix hardware.

COORDINATION WITH RELATED RESEARCH EFFORTS

     Related work has been going on at TVA's Division of Forestry, Fish-
eries, and Wildlife and at Oak Ridge National Laboratory.  Since the
beginning of this project, efforts have been made  to coordinate with these
groups to share knowledge as much as possible.  Discussions were held with
staff at ORNL, who are developing a parallel site  screening method, which
eventually led to the sharing of an ORNL-employed  planning intern with TVA
for work on this project.

     Meetings were also held with staff at TVA's Division of Forestry,
Fisheries, and Wildlife (DFFW).  This group has developed the use of a
batch-processed program, IMGRID, which will display geographically based
information by means of a grid system  (based on latitude and longitude)
superimposed on the study area.  After investigation of the feasibility
of using this software and DFFW's expertise, it was determined that IMGRID-
a batch, grid system—would not be compatible with the interactive
Tektronix equipment, which lends itself to polygon (rather than grid) data
storage.

CAPABILITIES OF THE COMPUTER SYSTEM

Statistical Analysis

     All the basic statistical analyses that were  needed were done on the
CRT terminal using simple FORTRAN programs or the  SAS packaged program.
This packaged program, designed for use by social  scientists, was quite
adequate for our purposes.  It would probably be adequate for any of the
basic analyses that planners in such a project are likely to need.

-------
                              C-14
Graphic Display

     The graphic display programs were developed by TVA's Division of
Environmental Planning and further refined by ND&RS, which provided the
principal programmer on the site screening project.  Three basic programs
were developed:  mapping of the counties, Kiviat diagrams, and scatter
diagrams.

Mapping—

     The mapping program draws the county boundaries within the study
area, identifies each county, and enables one to place a value or symbol
in the center of each county.  One county or a group of counties may be
drawn rather than the entire set.  The counties may be drawn at different
scales, allowing one county or portion of the map to be enlarged.  As
with all data handled in the CRT system, these maps are drawn and displayed
on the CRT with the option of producing a hard copy of each output.

Kiviat Diagrams—

     The Kiviat diagram (Figure C.2) is a graphic display method developed
for use in this study by TVA's Division of Environmental Planning.  The aim
of the diagram is to show how a single county ranks or scores on several
characteristics (this example is based on ranks).  In this example, six
indicators are shown, one on each of six axes within a circle.  The seventh
and eighth axes (at 90° and 180°) are not used; this gives the characteristic
"butterfly" shape to a diagram of the county that ranks high on all indi-
cators.  If eight indicators were used, all eight axes would be used to
form an octagon; similarly, four, five, or six axes could be used with or
without null axes and the polygon formed would vary accordingly.

     The distance from the center along each axis represents the rank (or
score) of the county on that characteristic.  In our example, the county
with the highest percentage of its population living in urban areas would
intersect that axis at the perimeter of the circle.  The middle-ranked
county would be halfway between the center and that perimeter, and the
lowest-ranked county would be closest to the center.  Thus, a county that
was ranked first on all six indicators in the diagram would show a perfect
butterfly shape.  The comparative quality of this graphics tool is demon-
strated in the several Kiviat diagrams shown in Figure C.3, which clearly
illustrates that Anderson and Hamilton Counties rank substantially higher
than, for instance, Hancock or Claiborne Counties on most of the indicators.

Scatter Diagrams—

     A graphics program was developed to draw scatter diagrams of pairs
of indicators.  As discussed earlier, these were produced for pairs whose
correlation coefficients were unexpected or otherwise of interest.  A
sample scatter diagram is shown in Figure C.4.

-------
     X  POPULATION LIVING  IN
          URBAN PLACES 
HOUSING VACANCY  C19705
            AVERAGE TEACHER'S
                SALARY
PERSONS  LJVING  IN  SAME
   HOUSE  OVER LAST  S
           YEARS
      HOUSING BUILT BEFORE
                1950  CX:>
 I
h-»
Cn
HEALTH  Cphyelclan* par 1000 p*r«on«>
       Figure C.2.  Sample of socioecoaomic indicators displayed as a Kiviat diagram.

-------
    ANDERSON
HAMILTON
                                                     HANCOCK
                                                HAWKINS
    CLAIBORNE
                              MCMINN
Figure C.3.  Butterfly configuration of socioeconomic indicators  for  various  counties.

-------
                         C-17

aa-
UJ
U 20 —
UJ
1-
V)
H 18-
CL
D
Q.
16-


A
— A
A A
—A A
A
A A A A
A A ^
— AA A A
,^
— A
A
1 1
iiiiiiiiiiiiii it''
  7000
8000        9000        10000
    AVERAGE  TEACHER'S
          SALARY
1100C
Figure C.4.  Sample scatter diagram.

-------
                              C-18
 INDICES

 Indicators

      Several  indices that were developed for testing combined indicators
 so  that  each  index covered a community system:

 Community Service Index--

      1.   Recreation acres (exclusive of Federal and State lands) per
          capita
      2.   Police expenditures per capita
      3.   Percent of population served by public water and sewer systems
      4.   Fire service indicator (to be added)
      5.   Physicians per 100 population

 Planning Index--

      1.   Percent of population living in jurisdictions served by planning
          commission
      2.   Percent of population living in jurisdictions having a compre-
          hensive plan
      3.   Percent of population living in jurisdictions having a zoning
          ordinance
      4.   Percent of population living in jurisdictions having subdivision
          regulations

Education Index—

      1.   Expenditure per pupil
      2.   Percent of pupils in overcrowded classrooms*
      3.   Average teacher salary
      4.   Number of pupils per teacher*

Urbanization and Growth Index—

      1.   Percent of county population living in urban places
      2.   Housing vacancy rate                j.
      3.   Percent of housing built before 1950~
      4.   Percent change in population (1960-1970)
      5.   Percent change in population (1970-1990)
      6.   Labor supply indicator (to be added)

Economic Need Index—

      1.   Employment rate     ^
      2.   Median family income"
      3.   Percent of population receiving welfare          ^
      4.   Percent of population in 20- to 44-year age group
*
 Indicators for which a low score rather than a high score is most favor-
 able.  Such indicators were subtracted rather than added in computation
 of the index.

-------
                              C-19


      On all indices, the higher a county's score, the better suited it
would be for the site of a large power facility.  However, the first
five indices are directed at measuring its capacity to absorb socio-
economic impacts, whereas the sixth index is designed to measure the
county's need for the potential benefits of such a facility.   Most
counties that have high capacity to absorb impacts do not have great
economic need, and vice versa.  Therefore, the economic need index was
intended to be viewed separately from the five capacity indices.

      A composite suitability index was developed to include the scores
on all indices.  However, inclusion of the need index only dilutes the
scores of all counties, and much information is lost.  Therefore, it was
decided to use a two-step process originally outlined in the site screen-
ing procedure used during the first year of work.  First, counties
received a composite capacity score based on their scores on the four
capacity indices.  Then, economic need indices of these counties with
the highest capacities were used to rank the high-capacity counties
according to their degree of economic need.


Weighting

      Some functional areas of  community life are directly or severely
affected by the  construction of a power facility, whereas others may
not be as seriously affected.   Furthermore, adverse effects on different
functional areas will have impacts of differing severity on local resi-
dents.  Finally, the indices are more or less arbitrary in number and
scope—the indicators could logically be grouped in ways other than that
which we have selected.  Health and education,  for instance, could be
combined or both included under "public services."

      Therefore, mere summation of the five capacity indices is not suffi-
cient to obtain  a composite index.  This would weight all indices equally,
and there is little basis for doing so.  A weighting scheme is an essential
element in the development of the indices.

      Weighting  of the indicators composing each index is important for
the same reasons.  Some of the  indicators are more significant than others,
some cover a broader scope, and the number of indicators in any given area
depends more on  the availability of good data than on some organic
definition of a  functional system.

      Twjo independent approaches to weighting were followed:  (1) three
professional planners from TVA's Division of Navigation Development and
Regional Studies, familiar with the study area, were asked to rank the
22 counties according to their  ability to absorb socioeconomic impacts;
(2) a panel of "experts" made up of eight planners and economists par-
ticipated in a work session in  which weights were agreed on for all
indicators and indices.  During this session, indicators were ranked by
participating individuals, and  concensus was reached after several rounds
of voting and discussion.

-------
                              C-20
PRODUCING THE INDICES

     The indices were produced by means of FORTRAN programs to compute each
index described previously.   To do so,  all scores were first standardized
(transformed into "z=scores").  The index formulas, which were the summa-
tion of these standard scores, were calculated with and without the weights
described above.  The weighted and unweighted indexes are as follows:

     Unweighted Education Index

         EDIDX = V4 - V5 + V6 - V7;

     Weighted Education Index

         EDIDX = (6 x V4) -  (8 x V5) +  (5 x V6) - (8 x V7);

     where  V4 = expenditures per pupil (standardized),

            V5 = overcrowding (standardized),

            V6 = average teacher salaries (standardized),

            V7 = pupils per  teacher (standardized).

     Unweighted Community Service Index

         CSIDX = VI + V3 + V23 - V24 -  V2;

     Weighted Community Service Index

         CSIDX = (2 x V21) + (6 x V2)  + [4 x fire (no number yet)] +
                  (8 x V23)  + (8 x V24) + (6 x V2);

     where VI  = recreation  areas per capita,

           V2  = police expenditures per capita,

           V23 = percent of  population  served by public water,

           V24 = percent of  population  served by public sewers,

           V2  = physicians  per 1000 population.

     Unweighted Planning Index

        PLNIDX = V8 + V9 + V10 + Vll;

-------
                         C-21
Weighted Planning Index

   PLNIDX = (6 x V8) + (3 x V9) + (8 x V10) + (5 x Vll);

where  V8 = percent of population living in jurisdiction served by
            a planning commission,

       V9 = percent of population living in jurisdiction having a
            comprehensive plan,

      V10 = percent of population living in jurisdiction having a
            zoning ordinance,

      Vll = percent of population living in jurisdiction having
            subdivision  regulations.

Unweighted Urbanization  and Growth  Index

    UGIDX = V13 X V14 -  V15 + V16 + V17 +  labor supply (no number);

Weighted Urbanization and Growth  Index

    UGIDX = (8 x V13) +  (2 x V14) - (3 x V15) + (5 x V16) +
            (6 x V17) +  (2 x labor  supply);

where V13 = county population living  in urban places,

      V14 = housing vacancy rate,

      V15 = percent of housing built  before  1950,

      V16 = percent change in population  (1960-1970),

      V17 = percent change in population  (1970-1990).

Labor supply  indicator is to be added.

Unweighted Economic Need Index

    ENIDX = V18  - V19 +  V20 - V21;

Weighted Economic Need Index

    ENIDX =  (4 x V18)    (8 x V19) + (2 x V20) - (5 x V21);

where V18 = unemployment,

      V19 = median  income,

-------
                              C-22


           V20 = welfare recipients,

           V21 = population 20 to 44 years old.

     In addition to computing the individual indices, these two routines
also computed a composite suitability index.  In contrast to the two-step
composite capacity determination described before, these computer indices
included the economic need index.  They were simply sums of the standardized
scores on each individual index.  After these calculations were made, it was
decided to return to the two-step process, eliminating economic need from
the composite index.  This second round of composite capacity indices was
computed as follows:

         CCIDX = sum of weighted standard scores in community service,
                 planning, health, education, and growth absorption indices;

     Weighted Composite Capacity Index

         CCIDX = (8 x CSIDX) + (3 x PLNIDX) + (5 x EDIDX) + (9 x UGIDX).

RESULTS OF THE TWO APPROACHES TO WEIGHTING

Judgments of Three Professional Planners

     The method of paired comparisons was used to obtain estimates of the
ability of the 22 counties to absorb socioeconomic impact.  All three
planners were familiar with the study area.  Each was interviewed indepen-
dently.  In the exercise, each county was paired with each other county and
presented to the judges, who were asked to indicate which member of the
pair had growth capacity or was more socioeconomically suitable as a poten-
tial site.   The judges were required to designate one of the pair as more
suitable.  No ties were permitted.

     Once the paired comparisons were complete, counties were ranked, and
the judges then reviewed results for consistency with their overall judg-
ments for county suitability.

          Spearman's Rank Order Correlation Statistics were used to analyze
the agreement between the three sets of ratings.  All the correlation
coefficients were above 0.95 and indicate nearly perfect agreement among the
judges.  The rankings of one of the judges are presented below:

                  High         Medium        Low
Hamilton
Washington
Anderson
Hamblen
Bradley
Greene
Blount

McMinn
Roane
Hawkins
Loudon
Jefferson
Rhea
Monroe
Cocke
Sevier
Claiborne
Union
Grainger
Meigs
Hancock
Polk


-------
                              C-23
Results  from the  Computation  of  the Composite Capacity Index

     Computer analysis  of  the data based on the weighting system described
above resulted  in the following  rankings of the 22 counties in terms of
socioeconomic capacity:

                   High         Medium        Low
               Hamilton      Roane        Cocke
Hamblen
Anderson
Bradley
Washington
Blount
McMinn

Greene
Loudon
Jefferson
Sevier
Rhea
Polk
Monroe
Hawkins
Claiborne
Meigs
Grainger
Union
Hancock

     The economic need  index  resulted  in  the  following rankings:

                  High         Medium         Low

               Hancock        Grainger     Hawkins
               Claiborne      Roane        Anderson
               Cocke          Polk         Blount
               Monroe         McMinn       Washington
               Meigs          Greene       Hamilton
               Rhea           Sevier       Hamblen
               Union          Jefferson   Bradley
                              Loudon

     It should be noted that  none of the  counties that are in the high
group of economic need  was  in the high rank in terms of socioeconomic
capacity.  One county,  Rhea,  was in the top rank in terms of need and
in the middle rank in terms of capacity.  Roane and McMinn counties are
in the upper half in terms  of both need and capacity.  In all other
instances, the greater  the  county's capacity, the lower was its need.

THE KIVIAT DIAGRAMS

     A Kiviat diagram was produced for each county, showing its standing
on all the individual indices in comparison with the other counties.  These
were done by (1) using  index  scores computed  from their rank orders on
weighted indices and (2) using index scores computed from their rank orders
on unweighted indices.  Diagrams based on weighted data are illustrated in
Figures C.5 and C.6.

     The raw scores were considered more suitable as a base for the indices
because of the numerous problems in using ranks.  The Kiviat diagrams of the
indices are not programmed  in the form of butterflies, as was originally done.
The actual form depends on  the number  of variables (indices) to be displayed
and whether null axes (forming the butterfly  "body") are used.  However, the
interpretation is the same:   The closer the figure is to a square, in this
case, the better the county.

-------
    ANDERSON
                           BLOUNT
                                                 BRADLEY
                                                                      CLAXBORNE
     COCKE
                           GRAINSER
                                                  QREENE
                                                                       HAMBLEN
    HAMILTON
                           HANCOCK
                                                 HAVflCZNS
                                                                       JEFFERSON
U6.   Community Services/
     Facilities Index

1*7.   Planning/Public
     Administration. Index

1»8.   Education Index

Up.   Urbanization/Grovth
     Index

50.   Economic .teed Index

52.   Composite Capacity
     Index
                                                                                                                                     O
                                                                                                                                     i
Figure  C.5.~ Kiviat  diagrams  based on six weighted indices—raw data.

-------
           LOUDON
MCMINN
                                                       MEISS
                                            MONROE
                                                   i    / ^  '  \  \
                                                   i-f-x-V-l
                                                                                                       o
                                                                                                        i
                                                                                                       ho
           POLK
 RHEA
ROANE
                                                                             SEVIER
           UNION
                               WASHINGTON
                          U6.  Community Services/
                                Facilities Index

                          UT.  Planning/Public
                                Administration  Index

                          u6.  Education Index

                          49-  Urbanization/Growth
                                Index

                          !; 0.  EC or: o:r: c \~,e e d I.'i tie x
Figure C.5.   (continued)

-------
    ANDERSON
                          BLOUNT
                                                 BRADLEY
                                                                      CLAIBORNE
                                                                                                                                    O
                                                                                                                                    I
                                                                                                                                    ro
     COCKE
                          BRAINBER
                                                 GREENE
                                                                       HAM8LEN
    HAMILTON
                           HANCOCK
                                                 HAWKINS
                                                                     JEFFERSON
                                                                                      k6.   Community Services/
                                                                                            Facilities  Ir.dex

                                                                                      147.   Planning/Public
                                                                                            Administration Index

                                                                                      l»8.   Education Index

                                                                                      1*9.   Urbanization/Growth
                                                                                            Index
Figure  C.6.   Kiviat  diagrams  based on four weighted indices—raw data.

-------
     LOUDON
MCKENN
MEIGS
                                                                             MONROE
      POLK
                                                                                                           i
                                                                                                           to
                              RHEA
                                                     ROANE
                                               SEVIER
     UNION
                           WASHINGTON
                          h6.  Community Services/
                                Facilities Index

                          ^T.  Planning/Public
                                Administration Index

                          ^8.  Education Index

                          i<9.  Urbanization/Growth
                                Index
Figure C.6.  (continued)

-------
                              C-28
     Examination of a preliminary set of Kiviat diagrams helped to establish
the need for dealing with economic need factors separately.  As can be seen
in the diagrams in Figure C.5, capacity indices and the economic need index
have a strong inverse relationship.  For instance, compare the patterns of
Anderson, Blount, and Hamilton (high-capacity counties) with Cocke and
Hancock (high-need counties).  In no case does .a relatively high-capacity
county also have a high economic need, although the moderate-capacity counties
(such as McMinn, Jefferson, and Roane) have the most balanced hexagonal patterns

     A final set of Kiviat diagrams is shown in Figure C.6.  These are based
on the four composite indexes—education, community services, planning, and
urbanization.  In these diagrams, the closer the shape is to a square that
touches the circumference, the greater the capacity.

                      RESULTS AND CONCLUSIONS

RESULTS

Refinements

     Data are yet to be collected for two indicators and for two counties
that were not originally included in the test area.  The indicators are
fire service (the State insurance fire ratings for each fire department,
aggregated by county) and labor force availability in each county for
appropriate occupational categories (as collected by TVA's Human Resources
Division).  Data on all indicators have been collected for Knox and Sullivan
Counties, which were added to the test area.  The new data were added to
the computer data sets, but have yet to be included in the final run of the
site screening exercise.

Site Screening Test

     The site screening test has five steps.

1.   Calculate functional system indices.  This was done by using the
     weighted formulas developed after the weighting session.  Each county
     was given a score for each of the five functional system indices
     (including economic need).  This task involves a very simple statis-
     tical computer program.

2.   Calculate composite capacity index.  The scores of each county on
     the four capacity indices (excluding the fifth index, economic need)
     were used to calculate the composite capacity index.  The formula was
     developed by using the weights to be developed by the staff and con-
     sultants.  Thus, each county will have a computer index score that
     is a weighted sum of the four indices.

3.   Produce Kiviat diagrams.  One set of Kiviat diagrams was produced
     with the four capacity indices.  Another set with those four indices
     plus the composite suitability and economic need indices was also pro-
     duced.  These were used as a supplementary method of making quick
     comparisons among the area counties.

-------
                              C-29
4.   Produce composite capacity maps.  This is the one step of the site
     screening test that had to be done primarily by hand.   The intent
     was to achieve a map showing counties shaded to varying degrees,
     depending on their relative suitability as project sites.  At the
     present time, there is no efficient, low-cost way to draw such a  map
     using the Tektronix equipment.  Therefore, hand drafting techniques
     are needed for the production of such maps (Figures C.7 and C.8).

     The index scores calculated in  steps 1 and 2 were grouped into thirds—
     the top one-third of the counties, next one-third, and lowest one-third.
     Each third was assigned a pattern.  They may be designated high capacity,
     average capacity, and  low capacity for absorbing socioeconomic impacts.
     One map was prepared for the overall capacity index, and another was
     prepared for the economic need  index.

     For rapid preparation  of these  maps, the computer-generated map of the
     study area was used as the base, with shading added by use of zip-a-
     tone patterns.

     These maps, particularly the  composite capacity map, will probably be
     the most useful of all site screening outputs for quickly evaluating
     the relative capacities of candidate  counties.

5.   Select most needy counties having  good capacity.  This may be accom-
     plished in two ways, both of  which were  used.  First,  the maps prepared
     in step 4 were used to identify those counties that fall in  the  top  two
     categories on both the composite capacity  and economic need  indices.
     This group of counties can then be designated the higher-priority  areas
     for power plant sites.

     A  second method is to  select  the high-capacity  (or  high  and  above
     average capacity) counties and  arrange them  according to their scores
     on the economic need  index.   This  produced a rank ordering of the
     half-dozen or dozen counties  according to  need,  which could  be used
     to help set  socioeconomic siting priorities.

CONCLUSIONS

     The  site  screening method developed during the  course of this  study
and described  in  this  report  has  several  advantages  over previous, more
intuitive methods.

     The  computer processing  of  data, particularly in the  formation  of
indices,  adds  greatly  to  the  amount  of information that  can reasonably be
handled!  Thus, planning judgments can be made on the basis of much more
complete  information.  The method  does require the collection of  many data
on a wide variety of subjects, but all are readily available and  most are
in published  form.

-------
                                                                  MEDIUM
                                                                  LOW
                                                                                                        o
        _o/ EXCLUDED FROM ANALYSIS
Figure C.7.   Overall capacity index.

-------
                                                                  MEDIUM
                                                                  LOW
                                                                                                      o
                                                                                                      OJ
         EXCLUDED FROM ANALYSIS
Figure C.8.  Economic need  index.

-------
                                 TECHNICAL REPORT DATA
                          /Please read Instructions on the /-mm1 before completing)
REPORT NO.
TITLE AND SUBTITLE
APPLICATIONS OF COMPUTER GRAPHICS TO INTEGRATED
ENVIRONMENTAL ASSESSMENTS OF ENERGY SYSTEMS
                                                         3. RECIPIENT'S ACCESSION NO.
                                                         5. REPORT DATE

                                                           August  1978	
                                                         6. PERFORMING ORGANIZATION CODE
AUTHOR(S)
                                                         8. PERFORMING ORGANIZATION REPORT NO.
Malcolm C.  Babb  and Harrison R. Hickey, Jr.
                                                            TVA/EP-78/10
PERFORMING ORGANIZATION NAME AND ADDRESS
Division of  Environmental Planning
Tennessee Valley Authority
Chattanooga,  TN  37401
                                                         10. PROGRAM ELEMENT NO.

                                                              INE 624C
                                                         11. CONTRACT/GRANT NO.
                                                                 79 BW
 SPONSORING AGENCY NAME AND ADDRESS
   U.S. Environmental Protection Agency
   Office  of Research & Development
   Office  of Energy,  Minerals &  Industry
   Washington,  D.C.   20460
                                                         13. TYPE OF REPORT AND PERIOD COVERED
                                                            Milestone 1976-77	
                                                         14. SPONSORING AGENCY CODE

                                                               EPA-ORD
 SUPPLEMENTARY NOTES
   This project is part of the EPA-planned and coordinated Federal Interagency
   Energy/Environment  R&D  Program.
 ABSTRACT
This.report  summarizes the first two years of research  designed to demonstrate
applications of computer graphics to environmental  analyses  associated with
the evaluation of impacts from development of conventional energy systems.
The work  emphasizes the use of storage-tube computer  graphics technology as
a means for  improving the interaction between the engineer-scientist and the
power of  the computer.  Computer graphics is also shown to be an effective
medium for summarizing and communicating information  about the environment
and pollution control alternatives to technical  specialists, managers, and
the public.   Also, many techniques of analysis previously considered impractical
can now be conducted on a routine basis.  Applications  to several fields of
analysis  are described in detail, including air  quality,  water quality, radio-
logical hygiene, industrial hygiene, scoioeconomics,  and  data facilities siting
with the  use of geographically referenced data.
           (Circle One or More)
                              KEY WORDS AND DOCUMENT ANALYSIS
                DESCRIPTORS
       Environments
       Energy Conversion

     Computers,  Computer Graphic
                                            b.IDENTIFIERS/OPEN ENDED TERMS
                                              Integrated Assessment
                                                                      c.  COSATI 1 icId/Group
                                                                           10A
3. DISTRIBUTION STATEMENT

     Release to Public
                                            19. SECURITY CLASS (This Report)
                                              Unclassified	
                                            20. SECURITY CLASS ,'Tiii.t page)

                                              Unclassified	
                                                                       21. NO. OF PAGES
       138
22. PRICE
>A Form 2220-1 (9-73)

-------